HOME (main menu)

TONY CAMPBELL articles

PORTOLAN CHARTS menu

INDEX

complete SITEMAP

What the site is ABOUT

WHAT'S NEW

For at least a century and a half attempts have been made to explain the origin of medieval portolan charts but no sustainable answer has been found. In the absence of solid documentary evidence, clues, deduction, analogy and intuition have had to be used instead. The revisionist conclusions that follow are the result of a lengthy analysis.

Portolan charts are characterised by unprecedented geometric precision and newly invented conventions, among which are an underlying mesh of direction lines, place-names restricted to the coast, empty seas, intentional generalisation of the coastline, and coded markings for navigational dangers. Furthermore, the very concept of a chart for marine navigation was itself new. The area covered by the oldest surviving marine chart, the Carte Pisane (c.1270), namely the Mediterranean, Black Sea and sections of the Atlantic coasts, would remain the norm for the next two centuries.

Before the appearance of the portolan chart an open-sea (pelagic) mariner, from any early period, would have needed a mental map of at least part of the Mediterranean, allied to the ability (prior to the invention of the magnetic needle) to determine direction, by means of a mental wind compass – for which there is indeed documentation. He would not have set out to sea otherwise. The mental maps would have been formed out of an expanding network of individual courses, memorised during a pilot’s life. Since many voyages would have involved ports around the sea’s periphery those, collectively, could have served to define the Mediterranean’s limits. Based on a collaborative pooling of those mental maps it would have been possible to ‘download’ the whole network as a single graphic design. Once the Mediterranean was visible as a whole, the conflicting estimates of a place’s position could be reconciled and omissions rectified.

Circumstantial evidence places the birth of the portolan chart in an unidentified west-coast Italian port, between 1154 and 1204. It was clearly designed by and for pelagic sailors and there is no evidence of institutional involvement or external financing. As an object of commerce, the charts had to be worth more to the sailor more than the cost of their hand-production. Since most of the charts’ users would have been illiterate, they carry no instructions for use, which had to be supplied orally instead.

No mathematical calculations or measurements were involved. In essence, the portolan chart was not a geographic or cartographic construct but a memorised experience transposed into a geometrically consistent graphic form, with the coastal outlines and toponymy added afterwards.

The essay starts with a summary critique of previous explanations for the origin of the portolan chart. None had managed to provide a comprehensive answer or one that is historically plausible, which helps explain why none has proved generally persuasive. Rather than devoting much time to their limitations – although their arguments are challenged at the appropriate points – it seems more profitable to set out instead what it is hoped is a feasible alternative thesis. Inevitably, it must remain scientifically unproven and partly speculative, but it does incorporate what little hard evidence there is, and tries to match that with the (sketchy) understanding of early navigational practices, married to what is known of human needs and motivation.

A forceful argument will be made in support of the statement that there are no identified antecedents for the portolan charts. In particular, it is shown that Muhammad al-Idrīsī’s Charta Rogeriana of 1154 could have played no part in the origins story, whether as model or imitation [Powerpoint, Slide 1]. ¹  Instead, the range of unique innovations seen in portolan charts (which, significantly, continue throughout the early period) makes it almost impossible that they could have had any predecessor.

The essay then states its central claim: that the charts must have derived, in some fashion, from the collective navigational memory of Mediterranean mariners. Open-sea sailing – as distinct from voyages that hugged an always-visible coastline – had as its first requirement that the pilot and helmsman knew where they were heading, before they left port. Proceeding without that knowledge would be more than foolhardy, verging on suicidal. That is at the heart of this thesis. The one-sentence foundation for this theory can be summarised thus: medieval pilots must have had a clear idea of the route to their destination, coupled with a reliable inbuilt mental wind compass that could envision the required bearing at all times.

Historians naturally seek out and value textual records. In this instance, the oldest surviving chart (the Carte PisaneAccepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118)) was certainly preceded by one pilot book (the Liber de existencia riveriarum et forma maris nostri Mediterranei – on which see a brief description in section E.1a), with a second (Lo compasso de navegare – for details of which see E.1b) being perhaps its approximate contemporary. The inter-relationship (if there is any) of those three unique, unsigned documents – none clearly datable – continues to be much discussed in the context of the ‘chicken and egg’ question of their respective priority. The charts’ supposed period of origin (now that the best estimate points to the late 12th century) is likely to have roughly coincided with the first appearance of a magnetic compass. Again, the historical record is unclear but an attempt is made to assess what significance (if any) the introduction of that navigational aid might have had for the genesis of the portolan chart. What would, however, have been of central importance was the wind compass held in a pilot’s memory. The discovery that the structure of this vital aid was neither described by contemporaries (apparently) nor had been analysed in any detail by previous historians, led to a lengthy digression co-authored with Roel Nicolai (see The mental wind compass of the medieval Mediterranean: the rediscovery of its structure and interpretations of its use, which has already been published separately).

The topics covered in this essay are listed in a separate hyperlinked MENU. Part 1 will delve into the charts’ prehistory to try to piece together the likely developmental stages that preceded the oldest survivor. ²  Spatial cognition, and mental maps in general, will be considered, along with medieval navigation and the charts’ possible (but here discounted) antecedents. Part 2 elaborates on the hypothetical ‘downloading’ of the mental map of the Mediterranean, while discussing, inter alia, the charts’ main features and purposes.

Recent (but pre-electronic) marine charts display almost all that a navigator needs, when used in conjunction with a pilot guide. The medieval chart, on the other hand, seems to have stood alone without secondary support. Or to be more exact, the essential elucidation must have come from sources little-discussed by historians, namely orally-transmitted instructions and the memorised experience of previous voyages. Orality and memory have a much larger place in portolan-chart history than literacy. Not only will it be argued here that the charts had no textual antecedent at all, but also that the portolani (Singular ‘portolano’ and more usually called ‘portolans’ – which can be confused with the charts) these were medieval guides to the coastal places around the Mediterranean, listing the distances between them. They also include open-sea (pelagic) routes, which had in the past been proposed as a possible source for the chart, must now be interpreted as depending on pelagic information from a marine chart for their very existence.

The essay is rounded off with a brief look at 16th-century surveys outside the Mediterranean, seeking any analogies that could have relevance for the prehistory of the portolan chart, followed by concluding remarks. An account of the essay’s methodology, summary conclusions and possibilities for future research are set out in separate webpages.

Constant reference is made to the following documents:

• Liber de existencia riveriarum et forma maris nostri Mediterranei (hereafter Liber)
  British Library, Cotton MS Domitian A XIII, ff. 114-129
  Editor: Patrick Gautier Dalché, Carte marine et portulan au XIIe siècle: le "Liber de existencia riveriarum et forma maris nostri Mediterranei" (Pise, circa 1200) (Rome: École française de Rome: distributor, Paris: Boccard, 1995) [available Online]
  This is the oldest surviving guide to the Mediterranean covering both the coasts and long-distance routes, probably dating from no later than c.1210. ³ 

• Lo compasso de navegare (hereafter Lo compasso). Berlin, Staatsbibliothek, MS Hamilton 396
  Editor: Alessandra Debanne, Lo compasso de navegare. Edizione del codice Hamilton 396 con commento linguistico e glossario (Brussels, etc.: Peter Lang for the Gruppo degli italianisti delle Università francofone del Belgio, 2011)
  Dated 1296 but probably based on a mid-13th-century model, this includes updating that might take the sole surviving exemplar into the 14th century. ⁴ 

• Carte Pisane. Paris, Bibliothèque nationale de France, Rés. G. B118
  The earliest extant portolan chart, thought to date from c.1270. For a thorough analysis see, ‘Does the Carte Pisane need to be redated?’, and, for the reasons behind its dating, the 16 September 2020 note at the head of the ‘Updates’.

“Among the research problems connected with the portolan charts, the question of their origin is perhaps the most intractable”. ⁵   Published almost thirty-five years ago, that statement still applies. Winston Churchill’s description of Russia’s intentions in 1939 as “a riddle, wrapped in a mystery, inside an enigma; but perhaps there is a key”, matches the present difficulty. It is hoped that Churchill’s optimistic coda will now prove to be equally apt in this case. Was the origin of the portolan chart hiding in plain sight all along?

For at least a century and a half, historians have struggled to reconcile the realistic outlines of portolan charts with their medieval context, since it was the first map covering a large area to be defined by its overall realism. The question usually asked was how that level of geometric precision could have been achieved with medieval instruments or mathematical expertise, and, above all, with the mind-set of the time? To suggest, as this essay does, the outlines of a possible solution may seem bold if not downright foolhardy. However, since nothing has plausibly replaced those earlier theories (now generally dismissed) which focused on a Classical or Islamic origin, or even a very ancient creation, the challenge needs to be faced. ⁶ 

There is no certain ‘history’ of the portolan charts prior to the earliest survivor, the Carte PisaneAccepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) , which, as was confirmed in 2016, and later by carbon testing, should be dated about 1270. For scans of the chart as it now is, and an 1852 facsimile made prior to the damage it suffered, see [Powerpoint, Slides 2 & 3, showing respectively the original and facsimile versions]. Prior to that is a prehistory almost as empty as the open sea, or with occasional half-seen features looming through the mist. Nobody recorded the charts’ birth. Neither the charts themselves nor any extant text throw direct light on their genesis or antecedents. Instead we have to rely on scattered written fragments: among those are passing references in later official records (usually emphasising their necessity at sea, without detailing how they were used), or comments about marketing matters, wills and legal disputes. ⁷  But nothing is said about their origin, purpose, on-board role, or the method used to reproduce the charts faithfully. If any pilgrim or crusader with an enquiring mind had sought information on those topics from a crew member, and then recorded it in their memorandum book, no trace of that has been found. The earliest landsmen’s descriptions of sailors consulting their charts date from the 15th century. ⁸ 

This leaves the charts as crucial witnesses to their own history. Most of what is reliably known about them derives from close analytical study of the oldest survivors, and particularly their toponymy.

In the absence of archival sources to provide part of the answers, a possible hypothesis will be outlined, echoing what Ramon Pujades expressed in his monumental survey of 2007, namely that "a historical solution must be sought to what others have tended to treat as a purely geographical-cartographical problem". ⁹ 

There are five interconnected questions needing to be addressed: when, where, by whom, how and why were the earliest portolan charts constructed? Cartometric analysis can certainly help with the technical component of the 'how' aspect, and explaining how the outlines made their way onto a map demands a joint historical and numerical answer, for which the results of cartometric analyses must be carefully considered. But the four other outstanding origin questions (when, where, by whom and why) are unlikely to be revealed through mathematics.

Instead, if there is any chance of determining where the portolan chart came from we must try to envisage the world out of which it emerged, the environment in which its creators and first users moved. The scant historical evidence must be sought out, both textual and cartographic, and when that runs out, we have to fall back on the constant verities of human nature and need, set into the specific context of medieval Mediterranean navigation. Ultimately, hypothesis, logic, common sense, intuition, analogy – none of which match the requirements usually demanded of historical evidence – will have to play an important role.

One aspect that must be squarely faced is the navigational capability of the medieval sailor. In a doctoral dissertation of 2014, Roel Nicolai concluded, at the end of a detailed cartometric investigation, that medieval mariners would not have had sufficient ability or equipment to have achieved the positional accuracy of the portolan charts: "The geodetic basis of the charts, expressed in their accuracy and their map projection, is far beyond the capabilities of the European Middle Ages". ¹⁰  That left him with the unanswered (and probably unanswerable) question: which era and place could have been responsible?

It appears to be the most likely scenario that Italian merchants acquired a body of ancient maps or charts through their trade with Constantinople ... The merit of the Italian mariner-traders would have been that they saw the potential of these maps, while the Byzantines, who preserved the heritage of antiquity but did nothing with it, may not have been greatly interested in them. It will have to be borne in mind that no person in the Middle Ages could have known how accurate these maps were. ¹¹ 

This essay calls for historical empathy, taking a more generous view of medieval aptitude and ingenuity, working on the assumption that people will always develop the skills they need for survival. It also argues that the experience of Mediterranean navigators would have been sufficient to supply the basis for the charts’ outlines, and that the generally accepted west coast of Italy should be retained as their place of origin, probably in the late 12th century. In other words, the central issue being proposed is not whether the early mariners had the geometrical ability to make a chart, but rather whether the level of their navigational skills, including direction-finding – which must have been commonplace among them for centuries – would have been sufficient for that purpose.

Prior to the emergence of the portolan chart, all navigational knowledge must have been memorised and shared orally. Hence the central task for any origins theory is to find a plausible way of explaining how the Mediterranean outlines made their way onto a chart in the first place. Significant research by Joaquim Alves Gaspar in 2019 has allowed us to reinterpret an important section of the oldest textual reference known, the LiberThe ’Liber de existencia riveriarum’, c.1210, written in Latin, is the oldest surviving guide to the Mediterranean, giving the distance and direction between places along the coast and across the open sea (Cotton MS Domitian A XIII, ff. 114r–129v). ¹²  Its statements of the distance and direction between paired places on the mainland coasts or islands, were previously considered to describe actual open-sea (pelagic) voyages. However, some were impossible, since they cross land, and that interpretation is no longer generally accepted. Instead – and this is the crucial point – it has become clear that the source for those statements was not direct sailed experience but rather measurements taken selectively from some kind of cartographic document. This confirmed that the portolan chart (or its antecedent) pre-dated the Liber, thus moving the focus away from a textual source towards an oral and graphic one.

Even more important, Gaspar was able to show that when the positions of all the paired termini at either end of those Liber statements were plotted out, a geometric diagram was created which defined the outlines of the entire Mediterranean with approximate accuracy. It is this (necessarily hypothetical) diagram – the result of what is here being termed ’inadvertent triangulation’An invented term for the hypothetical process by which pilots, before there was a marine chart, could have built up a mental network of the interrelationships between Mediterranean headlands, by means of observed directions and estimated distances. No mathematical calculations would have been involved – that, it is proposed, was the portolan chart’s source. So, the first question that has to be convincingly answered is this: what was the source of the pelagic information which would have been sufficient to build a wholly new cartographic image?

There are two other important linked questions that demand answers. First, given that their relatively small scale provided insufficient detail for the purposes of precise navigation, why were the charts conceived and created in the first place, and, second, why were they replicated for the use of paying customers for four centuries thereafter?

The ornamented charts that survive in disproportionate numbers from later periods were evidently not intended for use at sea. But if, as is still occasionally claimed, even unillustrated portolan charts served no navigational function either, then far more would have survived, instead of being worn out from overuse in inhospitable conditions. Furthermore, it is only once it is assumed the charts had a practical seafaring purpose that their peculiarities make sense.

While this essay is firmly focused on the charts’ hypothetical origins, it also seeks out evidence and meaning from the earliest documented periods, and especially from close examination of the oldest surviving charts themselves. Circumstantial evidence and the application of contextual logic offer a better chance of building up a picture, however tentative, of what the lost prototype chart might have looked like. There are some indications available, however, even if only by analogy. The Carte PisaneAccepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) and its distant descendants as late as 1700 share a cartographic language made up of so many common features, as well as consistent purposes, that there a strong argument for tentatively extrapolating those backwards into the charts’ prehistory.

New logical deduction allows the date of the portolan chart’s inception to be narrowed to a brief window. As is explained in more detail below, ¹³ no marine chart could have been circulating before 1154, because the fifteen years al-Idrīsī spent gathering data for his vast world map of that year the (Charta RogerianaA 3.5 metre wall-map of the world created, in Arabic, by Muhammad al-Idrīsī in Palermo, Sicily for the Norman king Roger II in 1154, and preserved, in the form of 70 sections, in several copies of a manuscript text) would certainly have included elements from a portolan chart if he had seen one. Based in Sicily, at the centre of the Mediterranean, it would have been impossible to avoid them. Conversely, working from the other direction, the new understanding that the Liber’s author was evidently using a marine chart for his pelagic measurements means that a functioning antecedent chart must have already been in place by the time that was written. The Liber’s sketchy details for the Black Sea (not taken from a chart) would fit in with a date after, but only shortly after, access was opened up to that in 1204. On the strength of those observations, the newly proposed date for the Liber is c.1210, and for the chart it was copying from, lacking the Black Sea, earlier than 1204.

In terms of the charts’ overall coverage, therefore, four phases can be distinguished: first, the core region, the Mediterranean, devised probably between 1154 and 1200; second, the eastwards extension to take in the Black Sea during the 13th century; third, the addition of the Atlantic coasts probably in the period 1300-1330; and then the steady recording of successive Portuguese exploratory voyages down the west coast of Africa in the 15th, leading to the voyages of discovery in the 16th.

There are intriguing parallels between the study of portolan chart origins and archaeology. What emerges from the soil might be the precise and complete outline of a building. Its purpose, say as a place of worship, might be broadly inferred from analogy with others elsewhere (if such existed). But who used it, what they believed, and what form their rituals took would probably remain elusive. In the Carte Pisane Accepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) we have the equivalent of that ancient building but little or no direct evidence as to how or why it was made. It did not come with a manual. With a leap of imagination, it might be possible to see the charts’ toponyms as akin to the archaeologists’ postholes, outlining a definite feature but requiring other evidence to establish their purpose – or in this case why those place-names were chosen.

The portolan chart is truly a cartographic masterpiece. The unprecedented step of setting down on parchment a precise maritime chart is comparable to replacing a quill-pen with a computer. The wide range of its cartographic conventions, almost none of which had been seen before, will be discussed individually. The word ‘invention’ seems appropriate, given that a number of the aspects of the chart’s cartographic language could, in a modern age, have been patented. It is easy to see cartographic innovations as unexceptional when they appear to be so sensible, and apparently timeless. But somebody had to have that ‘obvious’ idea, and then it had to be imitated. That the techniques introduced on portolan charts continued for so long is testimony to the clear and imaginative thinking of the unknown mariners who devised them.

They are the first maps of the Mediterranean and Black Sea that are fully recognisable to us, and what they offered seamen was unparalleled in its complexity and practical usefulness compared to anything made for landsmen. While merely a helpful guide for voyages that hugged the coast, the charts would speedily have become an indispensable aid for navigation when out at sea, where no land was in view to provide orientation. For an experienced pilot it would fill out the parts he did not already know, while a non-sailor would have gained access for the first time to the maps that mariners had always held in their heads.

Before starting to grapple with the origin question itself, we need to try and envisage the thinking and experience of medieval Mediterranean sailors prior to the portolan chart. It must first be understood that they were unlikely to have seen any cartographic image of their sea, and certainly not one that could have been of navigational relevance. Indeed, it can probably be assumed that mariners of the time had no clear concept of a drawn ‘map‘ of any kind. In devising a pragmatic working tool, the chart’s creators would have had to start from first principles. Their approach would have been free of ideology, imposed rules or precedent. What features were included and how they were treated depended on one thing only: how relevant they were for the charts’ purposes. That the cartographic conventions used on the oldest surviving chart cannot be found on any extant work from that time or earlier, is solid evidence that the portolan charts are sui generis. Their largest single innovation is undoubtedly an idea: the concept of the portolan chart, added to that the realisation that an animal skin, untrimmed, could be used to carry a map of the Mediterranean. In other words, what is special about the charts is their imaginative uniqueness. They were revolutionary not evolutionary. Paul Souriau’s verdict was that “Invention is not creation ‘out-of-the-blue’, but rather a combination of pre-existing ideas”. ¹⁴  There is little evidence of that in the case of the portolan charts even if they did emerge out of the deep blue sea.

Medieval navigators shared (broadly) the way they named places, how they described direction and how they measured distances – in ways that would be lost in later years and which we now struggle to replicate. Was it not logical that having already freely exchanged the details of pelagic courses and the mysteries of their wind compass they should pool their knowledge in a shared marine chart? After all, unlike other artisans, they were not in competition with one another; on the contrary, they were bonded by the common problems and threats they faced.

To put the achievements into context, the canvas under consideration concerns the world’s largest inland sea, the Mediterranean – with a width of 3,800 km and a north-south extent of 1,100 km, involving an overall coastline of some 22,000 km. To that should be added the 4,300 km coastal extent of the Black Sea, without considering those sections of the Atlantic coasts that were included on the standard chart.

This extended essay attempts to solve what may be the longest-lasting mystery in the history of cartography. ¹⁵  The conclusions it reaches run counter to existing theories that the charts’ source was text-based: whether some record of individual pelagic courses, or the littoral toponymic progression documented in the portolani(Singular ‘portolano’ and more usually called ‘portolans’ – which can be confused with the charts) these were medieval guides to the coastal places around the Mediterranean, listing the distances between them. They also include open-sea (pelagic) routes. Instead it will be argued, from the best indications available, that the origin should be identified in the joined-up network formed of the termini of those open-sea (pelagic) courses, which had been held from time immemorial in the memories of Mediterranean sailors.

There is an essential distinction that whereas the land divides (whether by language, culture, politics or topography) the sea unites. In medieval times, on terra firma, outside Latin-using circles, there was a babel of languages. Similarly, units of linear measurement had proliferated, even when they were notionally based on parts of the human body. Measurement would not come to be widely standardised until the metre was devised in Revolutionary France. At sea, on the other hand, such disparities were negated by a unifying environment made up of water (only very loosely divisible), the islands dropped into it, and the abutting coastlines.

In ‘Men of Sea. The making of an Identity’ one passage defines what sets sailors apart from others [note the timeless comment in the final sentence]:

The identity of the ancient sailors or men of sea according to our research is based on an open opposition to earth-based society and it is recognizable through the extremely dangerous conditions of their work and their isolation, their particular customs and their deep religiosity and superstitious attitudes, obviously encouraged because of the evident risks of their activities. These basic features were basically shared for every sailor from Antiquity onwards. ¹⁶ 

Life at sea did not just benefit from co-operation it must have largely depended upon it. Leaving aside declared enemies, it would be surprising if those involved in maritime trade had not developed a tradition of mutual assistance. Even if there was commercial rivalry, that would have taken second place to the sharing of information about optimal routes, navigational dangers, the presence of pirates, any nearby ship afflicted with plague, local weather conditions, and so on. They were, after all, engaged in an endless struggle in the face of shared difficulties and risks. In that, contradictory sense, "If land is cosmos, ordered, the sea or the ocean is chaos, unordered". ¹⁷  Why would long-distance pilots have held back potentially life-saving information from those who might in time reciprocate? ¹⁸ 

So, where social cohesion on terra firma had to contend with various fissiparous tendencies, the marine experience was very different. The Mediterranean was a shared sea and sailors were able to travel, in principle, to any part of it without hindrance. Its ports, along with the ships’ crews, housed a shifting population that was probably more heterogeneous than could be found anywhere else. Agreeing on a single unit to measure distance, sometimes across the featureless sea, must have been a precondition for the piloting of merchant vessels in the open sea. Hence the adoption of a ‘portolan mile’, even if those who used it seem not to have realised that there were regional differences in the way it was interpreted. ¹⁹  Seamen also needed to agree on consistency in the naming of places and natural features, as well as an understanding of the subdivisions in their shared wind compass. ²⁰ 

Besides those bedrocks of maritime lore there was a common cartographic language for the symbols on portolan charts that highlighted hazards and (admittedly, a slightly later development) the intentionally unrealistic shapes given to the islands in the Aegean and elsewhere as a mnemonic aid. When the location of the prime meridian at Greenwich was finally decided in 1884, replacing numerous nationally preferred alternatives, that was no more than an acknowledgement of the fact that most mariners were already using Greenwich, paralleling that same tendency towards maritime conformity that can be seen in the medieval period.

It is remarkable that, during a period of at least 150 years, there has been no plausible explanation as to precisely how the information shown on the portolan charts was first gathered and subsequently processed. Roel Nicolai’s impressively thorough doctoral thesis was forced into a negative conclusion: “The geodetic and cartographic origin of portolan charts does not lie in medieval Europe”, but he was unable to offer an alternative birthplace. ²¹   The explanation, and there has to be a rational one, cannot depend on some unidentified earlier civilisation.

Were the charts created out of a coastal survey, by stitching together regional surveys, or by joining up into a single system the open-sea (‘pelagic’) courses across the Mediterranean?

There seem to be four possible hypotheses:

(A) Joining up the statements in a portolano

Jonathan Lanman’s 1987 essay ²² purported to demonstrate that an overall chart could have been derived by plotting onto a map the stated directions and distances from one port to the next all the way round the Mediterranean (with its islands) and the Black Sea, as is described in Lo compasso’Lo compasso de navegare’, c.1260-96, written in Italian and apparently designed for sailors, is much more detailed than the earlier ’Liber’ (Staatsbibliothek, Berlin, MS Hamilton 396).

That must fail for one simple reason. While plausible outlines might be arrived at by those means for short stretches of coastline, the inevitable imprecision of both the stated distances (in miglia = miles) and the ‘wind compass’ directions would have made it impossible to build a total map of the Mediterranean in that way. ²³ 

Taking into account expected textual error and omissions in the sources used, this would have introduced steadily growing distortion into the broader picture. Such a method, carried out in sections, would never have reproduced the overall geometry of the Mediterranean in any usable form. There could have been no accurate matching of the north and south coasts, no reliable pairing of Spain’s features, for example with those of the Levant, and no understanding of the correct proportions of the Black Sea. As Joaquim Alves Gaspar concluded: “However recognizable in general terms the resultant outlines of the Mediterranean Sea, their accuracy and detail are hardly comparable with those of the charts they were intended to replicate”. ²⁴ 

It is true that no one, to my knowledge, has directly suggested such a systematic coastal survey but it seems to be generally assumed (other than by Roel Nicolai and proponents of an Islamic origin), ²⁵ that the portolan charts’ information is entirely medieval. It is just that nobody has been able to suggest how that might have been gathered and brought together in the form of a marine chart.

The contention that the portolan chart emerged from the joining up of a series of sub-charts of the basins and zones of the Mediterranean, which seems to be increasingly accepted, is discussed in detail later – and there critiqued in detail. ²⁶ 

In a 2023 article, Gregory McIntosh interprets the origin of the portolan chart as a post-compass invention devised by pilots entirely for navigational reasons and 'without performing any trigonometry': 'By conveniently combining onto a single piece of parchment his [the navigator's] courses and triangles ... the proto-portolan chart was born ... The portolan chart was the product of a long, evolving process of improvement and refinement of the tools and techniques of the medieval mariner ...'. {This paragraph added 24 June 2023}. ^26a 

Any search for plausible antecedents of the portolan chart needs to ask which of the cartographic models known to have been extant [though not necessarily accessible] at the end of the 12th century might have played a part. In what ways would different cartographic types, exemplified for example by the mappaemundi'World maps' in Latin (singular, mappamundi). Used to distinguish medieval maps of the world from those of Ptolemy and the portolan charts, the Peutinger Table, Matthew Paris’s itinerary maps, the Charta Rogeriana ²⁷ or regional maps and diagrams, have helped provide any kind of realistic picture of the Mediterranean and Black seas? ²⁸ [Powerpoint, Slides 1 & 5]. Furthermore, who, in 1200, would have needed, or been able to make use of, a realistic outline of the known world, with positional geometry as its primary concern?

Ptolemy’s astronomically determined co-ordinate geography might seem to be a suitable candidate (even assuming that it was available in Western Europe at that time via Islamic intermediaries, which is unlikely). The Geographia’s coordinates set out what was known in Alexandria in the second century CE and would not start to be updated until the Renaissance. In the period when the portolan chart seems to have emerged, cartographers had other priorities besides Euclidian accuracy. Indeed, despite the example of the portolan charts, with a few exceptions a regional cartography that used a measured scale and concentrated on the correct relative positioning of places – what might be termed ‘proto-scientific’ – is not seen for several centuries.

We should not have to rely on hypothetical antecedent maps, which are not only physically lost but also absent from the written record. If it is felt that such must have existed, how is it envisaged they were created and who would have drawn or used them? Those who support the claims of non-mariners as the charts’ progenitors need to identify who else, realistically, could have been involved. It is much more likely that the nascent portolan chart was ‘hidden in plain sight’ in mariners’ memories than that an invaluable Roman device was lost for 1000 years. Even if Greek mathematicians or Roman surveyors had turned their minds to the question of a marine chart, where would they have obtained the necessary information? Their periploi provide even less information than the medieval successors to those, the portolani (Singular ‘portolano’ and more usually called ‘portolans’ – which can be confused with the charts) these were medieval guides to the coastal places around the Mediterranean, listing the distances between them. They also include open-sea (pelagic) routes. Furthermore, the identical source question posed about the medieval charts would have to be asked in the case of a hypothetical Classical origin.

The iterative argument, which assumes that portolan charts must represent a continuation of one or more preceding cartographic genres, is not only contradicted by the lack of any such models but, even more compellingly, by the charts’ inclusion of a suite of unprecedented conventions, probably, in both number and originality, without parallel in the history of cartography. ²⁹ 

It cannot be said for certain that the portolan charts’ creator(s) had never seen a map of any kind. ³⁰  However, we can be sure that whoever drew the Carte PisaneAccepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) was unaware of any map attempting to cover seriously those sections of the North Atlantic coasts which he casually included: whether Ptolemy, the descendants of a Roman model typified by the Anglo-Saxon (or Cotton) map, ³¹  or al-Idrīsī’s Charta RogerianaA 3.5 metre wall- map of the world created, in Arabic, by Muhammad al-Idrīsī in Palermo, Sicily for the Norman king Roger II in 1154, and preserved, in the form of 70 sections, in several copies of a manuscript text. ³² The most logical explanation for that new cartographic language lies in the portolan charts’ clear focus on navigation – to the extent that almost everything else is omitted – confirming that they were created, de novo, for the use of navigators.

Assertion: Most sailors around 1200 were illiterate so why would a textual origin for the charts be likely?

Possibly because historians are used to dealing primarily with textual evidence, there seems to be a tendency – even perhaps an assumption – that, if the origin of portolan charts is ever discovered, it will take a written form, or at least have emerged from a literate group. Ramon Pujades articulated that viewpoint clearly, as in the following examples:

They were the products of the written culture of a specific period

The chronology of the advent and dissemination of nautical charts is totally inseparable from that of the rise of vernacular Romance languages to the category of vehicles of written expression

The information the [Liber’s] author had compiled to write his portolan and draw his mappamundi came from data on distances and directions amassed by a number of seamen-merchants-military men from the second half of the twelfth century who had recently learnt to read and write ³³

Ramon Pujades sets out his latest thinking on the origins issue in a major publication: Els mapamundis baixmedievals: del naixement del mapamundi híbrid a l’ocàs del mapamundi portolà / Late medieval world maps: from the birth of the hybrid to the demise of the portolan mappamundi, (bilingual, Catalan and English), Barcelona: Institut Cartogràfic i Geològic de Catalunya (ICGC), 2023 [but released online in March 2024]. The following extract (p.38, or 60 in the online version) is a useful summary from part of a single paragraph:

“The combination of geographic knowledge derived from commercial navigation and the growth of the literate sector of the seafaring population found in the Tuscan-Ligurian coastal arc from the second half of the 12th century onward provided the foundation allowing the appearance of the first partial ‘route guides’ of the Mediterranean – lists of distances and directions between the main ports and geographical features of those waters. These route guides were called portolani, a term that should be reserved exclusively for such and not be applied to portolan charts or nautical charts to avoid confusion. They compiled both sources of basic information needed for the new cartography, i.e. distances and directions. All that was left to do was to gather and organize those partial lists and fill in the blanks with the oral descriptions provided by experienced seafarers to turn them into a general description of the Mediterranean and, from there, make the leap of mental abstraction required to go from this written data to the graphic format of a map. In my opinion, this is where the spontaneous, direct contribution of the seafaring world stopped, however, and the scholarly world took over, consisting of those trained in the tradition of encyclopaedic compilation mentioned above, and who used Latin as a basic language of communication. This sector had the necessary geometric and technical knowledge and was used to working with the concept of the map or oecumenical diagram, lending it the mental capacity to envision the general layout of the three continents of the oecumene surrounding the Mediterranean Sea.”

Since the text of this book had entered the publication process in 2019 (and then met with Covid and other delays) Ramon Pujades could not have commented on my 2021 'Origins' essay even if he had wanted to. As a result, his hybrid hypothesis, of a joint inspiration involving both the ‘seafaring’ and ‘scholarly’ worlds, could not engage with my own mental-map thesis involving pelagic pilots alone. [This section added 6 October 2024].

Likewise, Patrick Gautier Dalché suggested that a pilot book which pre-dated the LiberThe ’Liber de existencia riveriarum’, c.1210, written in Latin, is the oldest surviving guide to the Mediterranean, giving the distance and direction between places along the coast and across the open sea (Cotton MS Domitian A XIII, ff. 114r–129v) would have been at least a partial source for the charts. ³⁴  The most likely textual source had at first appeared to be the Liber. However, recent re-interpretation of its pelagic statements ³⁵ has concluded, first, that those do not describe actual open-sea courses and, second, that the information they contained could only have come to the Liber’s author from a lost cartographic drawing, from which he had extracted those measurements. Thus the earliest relevant surviving text proves not only to have played no part in the charts’ invention, but also to post-date what must have been some kind of graphic antecedent to the portolan chart.

Besides there being no surviving textual source, there are no references to missing works of that nature either. Even if a supposedly relevant text was identified it would do no more than prompt a further question: what, in turn, was its source? It would be like the rainbow, not the crock of gold at its foot. Furthermore, it is argued here that there is no logical reason to suppose that there would have been any textual stage at all in the charts’ origin and early development. ³⁶  Instead, a diagram of open-sea courses based on sailors’ shared memories is posited as the underlying element, with coastlines and toponyms provided separately afterwards.

The portolan chart is a graphic document with a quasi-mathematical underpinning. No sign of a literary element has yet been identified in the charts’ DNA. That is surely no more than would be expected when the navigators’ milieu is taken into account. Sailors would probably have met, not in the study of a savant, let alone a monastic scriptorium, but on board a ship or in a quayside drinking establishment. Why is it logical to expect that a mental ‘map’ should have been downloaded into text before being re-envisioned in a graphic form which mirrored the original information? As the commentators on the 11th-century Islamic Book of Curiosities point out: “a transition from text to map is not as intuitive as it may seem to us today”. ³⁷ 

This chimaera of a missing urtext out of which the portolan chart emerged must be unmasked as an unfounded assumption (perhaps wish?) for the charts to have a documentary rather than mixed graphic and oral origin. ³⁸  The textual-origin argument ignores the evidence that much of the crucial information on the portolan charts could have come only from memorised experience. Nor could the charts’ navigational detail (for example the coded dangers) have been reproduced adequately in a text.

From an overview of early Islamic navigation, carried out by Yossef Rapoport and Emilie Savage-Smith, in connection with the 11th-century Book of Curiosities [Powerpoint, Slide 6], it has been established that there was plentiful, detailed information available to Arabic speakers and readers. ³⁹  Indeed, the book’s “recording of navigation data in a systematic way is unique and distinctive”. ⁴⁰  The portolan chart, despite its great originality, must defer in one overall respect to that Arabic manuscript, which has been justly described as the “first surviving example of a map looking at the coasts from the sea”. ⁴¹  However, realistic coastal outlines were intentionally excluded from the maps in The Book of Curiosities. What is notable is the mis-match between the text and the accompanying graphic representations, which were abstract and generalised. Indeed the work’s author explains that his marine maps were not intended to be “accurate representations”. ⁴² 

It is valid to ask why the Islamic world did not invent the portolan chart, given that in the Mediterranean, “a shared and contested space, the problems facing Muslim and Christian navigators were the same. Much of the maritime nomenclature was common across the boundaries of faith, and the technologies that Christian and Muslim mariners used were broadly similar”. ⁴³  Despite this, Rapoport and Savage-Smith rebut past claims to that effect: “In view of the evidence of the maritime maps of The Book of Curiosities, it seems highly unlikely that some early version of portolan maps was known in Fatimid Egypt. This is worth emphasizing because some have argued that portolan maps first emerged in the Islamic world”. ⁴⁴ 

That Italy rather than the eastern Mediterranean was the birthplace of the portolan chart might have been no more than accident. But it should not be assumed, from the overlapping navigational needs, that there were no major differences. One of the largest was the omission in The Book of Curiosities of references to hazards faced by shipping, yet that was certainly one of the main purposes of the portolan chart. Furthermore, the Islamic text’s concentration on bays, ⁴⁵ which presumably shows a concern for ports with their mercantile relevance, contrasts with the portolan charts’ focus on the feature of the greatest concern for navigators, namely headlands. ⁴⁶ 

A.3b. al-Idrīsī’s Charta Rogeriana of 1154

The map under consideration was produced in 1154 or a little later, by Muhammad al-Idrīsī, for Roger II, king of Sicily. It has come down to us in the form of seventy separated sections, which, if they had been joined up, would have had the proportion of a 3.5 metre-wide wall map. That masterpiece needs to be clearly distinguished from a much-reproduced small circular world map, whose connection to al-Idrīsī is tenuous and whose importance has been overstated.

The large map, which is here referred to as the Charta Rogeriana, is dispersed through the pages of a manuscript volume, Nuzhat al-mushtāq [The book of pleasant journeys into faraway lands]. At issue is the question of whether the Charta incorporates any of the conventions otherwise specific to the portolan chart. If that was the case it would indicate that the latter was in existence prior to 1154. Alternatively, since the Charta is the only surviving map from the period before 1200 that could conceivably have provided inspiration for the portolan chart’s creators, is there any evidence that they had borrowed from the Charta? ⁴⁷  In other words, which came first, and was either influenced by the other?

The attempt to answer that apparently simple question ran up against the problem that many of the details of the Charta’s creation and its later dissemination had to be teased out from conflicting and contradictory evidence. That analysis can be read in greater detail in a recently published web article. ⁴⁸  The main conclusions that confirm the lack of mutual influence between the Charta and the portolan chart are set out in the following sections, which contrast the content and purpose of each.

A.3b.1. Comparison between the content of the Charta and the portolan charts

The Charta’s place-names are at their densest around the coasts – which would fit in with Roger’s desire to know about 'seas and gulfs' – but they appear to refer to settlements only, not to natural features, though a few rivers and mountain ranges are identified. However, despite the Charta’s deserved reputation for innovation and thoroughness, it would have been of limited use to a sailor. It is true that what is considered one of the signature features of the portolan charts, namely the placement of the toponyms inside the coastline, is also found on the Charta, but there is a significant difference. Whereas the portolan chart’s toponymy winds its way round the coastline in an unbroken sequence, with the names placed at right angles to that – which meant that the user had to rotate the chart to see them the right way up – the Charta’s toponyms are generally placed horizontally, to be read from a single viewpoint like normal text. Furthermore, the Charta’s patterned sea indicates there was no intention of depicting any offshore dangers, which was one of the portolan charts’ primary functions. Not even the feared Skerki Bank, to the west of Sicily – and hence not far from Palermo, where the Charta was drawn – was included. ⁴⁹ 

There are a few similarities between the Charta and the Carte PisaneAccepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) , for instance in the coasts around Tunis. ⁵⁰  Nevertheless, it is difficult to see how al-Idrīsī’s own reported visits to the Maghrib could have had anything to do with those outlines. Moreover, the single most notable feature of the portolan charts – indeed their most distinctive coastal signature: the deep Gulf of Sirte – is only partially conveyed on the Charta where it is given a similar depth to that of two other, non-existent gulfs. Where the Charta is visibly more advanced than the Carte Pisane is in its portrayal of the Atlantic coasts, which lay outside the knowledge of that chart’s author.

1. The Ptolemaic image was formed explicitly from the astronomically calculated latitudes of hundreds of places, with their longitudes depending on reported distances. Relatively few of the place-names, around 13%, represent natural coastal features.⁵¹ 
2. The Charta’s unstated latitudes were derived from the seven equally spaced ‘climates’ of al-Khwārazmī.
3. By contrast, the portolan charts, as is claimed here, were based on a diagram made up of the termini of interconnecting pelagic courses across the Mediterranean. These would have been downloaded from navigators’ pre-existing mental maps, with subsequent reconciliation of the inevitable variations in the placement of the termini. They were marine charts and hence concerned almost exclusively with precisely-positioned ports, islands and coastal features.

A.3b.3. Why the portolan charts were not used by al-Idrīsī

Roger II – or more realistically al-Idrīsī himself – spent fifteen years interviewing those who passed through the capital, Palermo, asking them about the geography of the world’s regions. A number of those informants must have been sailors. Had al-Idrīsī been shown a portolan chart, we can be confident that its user would have vouched for its accuracy and hence relevance for Roger’s project. Yet there are only fragmentary signs in the Charta of input from sailors and certainly no traces of the portolan charts’ many unique features.

It is always dangerous to argue from the absence of information but, in this case it seems justifiable because it is inconceivable that al-Idrīsī’s meticulous and well-documented, information-gathering exercise would not have revealed any nascent portolan chart that might then have existed. My failure to identify any traces of borrowings from a marine chart on the Charta, ties in with the majority opinion among researchers today. Tarek Kahlaoui, for example, concludes that "it is extremely difficult to argue for the Idrisian use of early portolan maps". ⁵² 

The significance of this cartographic silence – which has not apparently been noted previously – provides robust support for a portolan chart origin date of no earlier than 1154 (when al-Idrīsī began his final version) and around 1158 when he completed it). ⁵³ 

The first and most obvious reason for this was the Charta’s inaccessibility. The 3.5-metre map was copied out in 70 sections, which were then distributed through an accompanying geographical text. The first time the full image is known to have been available was in a redrawing of 1928. Faced with the double impediments of a fragmented image and Arabic script, it is highly unlikely that anyone from outside the Islamic world would have seen the Charta or, if they had, would have been able to make sense of it – and certainly not at the time when the portolan chart was being developed. Hence, al-Idrīsī’s world map could not have served as any kind of inspiration for the portolan chart. ⁵⁴ 

The reverse question might be even more relevant: how could those navigating the Mediterranean before the compass not have had their own mental maps? “There has probably always been a mapping impulse in human consciousness”, as J.B. Harley wrote on the first page of the first volume of The History of Cartography, “and the mapping experience – involving the cognitive mapping of space – undoubtedly existed long before the physical artifacts we now call maps”. ⁵⁶ 

This section leans heavily on M.R. O’Connor’s, Wayfinding: the science and mystery of how humans navigate the world (2019) [Powerpoint, Slide 7]. ⁵⁷  The author, an American journalist, ranges widely and deeply over the subject of human wayfinding. She shares her first-hand experiences in the company of Inuit hunters in the white landscape of the Canadian Arctic, Aborigines wayfinding in the Australian desert, and Oceanic mariners travelling vast distances guided only by the stars and waves. All of that is interwoven with a survey of the fast-moving developments in brain function research, focusing on the neurological basis of spatial orientation, and buttressed by interviews with the leading figures in the field. The science is not always easy but she takes us patiently through it. Though it is not directly concerned with Mediterranean navigation, this overview is a useful way into the theory and practice of human wayfinding.

Research continues apace into this issue, as chronicled in an overview by Pablo Fernandez-Velasco & Hugo J. Spiers, 'Wayfinding across ocean and tundra: what traditional cultures teach us about navigation’. {This sentence added 9 November 2023.} ^57a 

Recent research into spatial orientation, as summarised by M.R. O’Connor, identifies two types:

The first is route knowledge, an ability to construct a sequence of points, landmarks, and perspectives that make up a path from one place to another. The traveller uses a string of memories of landmarks or viewpoints to recognize the correct sequence for getting from one place to another … The second strategy is called survey knowledge: the traveller organizes space into a stable, maplike framework, in which every point or landmark has a two-dimensional relationship to every other point ... Route knowledge relies on the traveler’s point of view and relationship to objects around them, what’s called egocentric perspective … Survey knowledge depends on what is called an allocentric perspective, a point of view that is objective, maplike … ⁵⁸ 

Clearly, pre-chart Mediterranean navigation would have routinely involved a combination of the ‘route’ and ‘survey’ strategies. Following a remembered or prescribed course formed out of visible waymarks, on the one hand, and navigating ‘blind’ in the open sea aided only by a mental map of the inter-relationships of places and features aided by a mental wind compass, on the other.

As Anita Devineni points out, the spatial map identified in the hippocampus may serve “primarily to encode memories: events are always associated with a particular place, so linking the place to the event may be a handy way of ‘indexing’ your memories”. ⁵⁹  Such an aptitude would have had great relevance for memorising a long sea voyage with its occasional sightings of land. ⁶⁰ 

B.2b. Is spatial cognition innate in humans?

Could we be born with spatial memory recall? There are those who have suggested this might be the case. “While languages and navigational strategies may vary from one culture to another”, wrote John Edward Huth, “the wiring of the brain to create languages, mental maps, and other skills appears to be an intrinsic feature that we’re born with. Navigational skill depends on the interplay between memory, perceptions, and the mental map”. ⁶¹  David Lubinski thought that spatial ability “may be the largest unknown, untapped source of human potential”, while Alfred Ardila noted that “for thousands of years, human survival depended on the correct interpretation of spatial signals, memory of places, calculation of distances, and so forth, and the human brain must have become adapted precisely to handle this kind of spatial information”. ⁶² 

Given the origins of human evolution, it would be unwise to rule out the possibility that we have spatial navigation skills that have some equivalence to those being identified via tests on animals. Edward Tolman’s celebrated 1948 article, ‘Cognitive maps in rats and men’, ⁶³ has been followed by other studies looking for parallels. Many animal abilities far exceed our own, whether in measuring time or maintaining direction over perhaps thousands of kilometers, but it is now thought likely that all species have some kind of a bio-compass that can respond to the earth’s geomagnetic field, even if, despite half a century looking for it, the evidence has not yet been found. ⁶⁴  As the Inuit historian, Ken MacRury, explained about hunters in the Canadian Arctic: “They wouldn’t get lost. And the dogs never got lost, never. Fifteen or twenty years ago, the old Inuit couldn’t believe when people started getting lost. They couldn’t believe it was possible.” ⁶⁵

The place in the brain where the mental map has to be looked for is the hippocampus.

B.2c. The hippocampus

The flurry of recent research into this area of the brain continues apace and doubtless in ten or twenty years’ time there will be far more certainty about the function of that and other areas associated with navigation and memory, for example the caudate nucleus. It is clear that the hippocampus is critical for both cognitive mapping and episodic memory. It can objectively represent the environment in three-dimensional space and allows spatial relationships to be inferred (e.g. by taking a short-cut).

The hippocampus is very malleable. It includes “head-direction cells, which discharge in relation to which way our head is pointed on the horizontal plane, and grid cells, which fire as we roam an environment and build a coordinate system for navigating”. ⁶⁶  Essentially, the brain links its map of where things happened to memories of what happened. The key to the growth of its grey matter is experience, not time, and it responds particularly to complex environments. Growth can effectively start as early as six years old, and it can be observed to expand in response to formal memory exercises. The central relevance of the hippocampus to the conjectured mental map of the medieval mariner is obvious.

B.3a. Definition of a ’Mental map’

What is a ‘mental map’? The definition given in Wikipedia is as follows: “A cognitive map is a spatial representation of the outside world that is kept within the mind, until an actual manifestation (usually, a drawing) of this perceived knowledge is generated, a mental map. Cognitive mapping is the implicit, mental mapping the explicit part of the same process”. ⁶⁷  Jeremy Black makes a similar distinction between ”'unmanifested mental maps' (those that exist only in the mind), 'manifested mental maps' (such as sand paintings and Australian aboriginals carvings), and printed maps". ⁶⁸ 

On the other hand, in 1987 one meaning of ‘mental map or mental cartography’ had been defined in The History of Cartography, as “an image of the environment held in the mind to aid wayfinding or spatial orientation … [which] groups apparently carry in their heads as mnemonic devices”. ⁶⁹  A possible current term might be ‘virtual map’ but, in the light of modern usage, that would be confusing and will be resisted here. Since this essay will discuss both the hypothetical content of sailors’ minds and its later manifestation in portolan charts, it would be unhelpfully pedantic to alternate Wikipedia’s distinction between ‘cognitive’ and ‘mental’ maps in those different contexts. ‘Mental map’ will therefore be used throughout.

Much could probably be learnt from the modern Vendée Globe, a single-handed, non-stop yacht race, in which participants sail around the world without the navigational aids taken for granted today. Earlier entrants were denied GPS and today they have to rely solely on sextants, paper maps and the stars; the magnetic compass is banned. A rather different example of traditional navigation is the story of the Hõkūle’a, a double-hulled Polynesian voyaging canoe, which covered nearly 50,000 nautical miles over a period of forty years, without the use of Western instruments, maps, or charts. ⁷⁰ 

We no longer have access to medieval mariners but might it be possible to mount a research project that scanned the activity and physical development in the brains of these oceanic navigators undergoing the Vendée Globe challenge? [Although the all-important memories of past routes in the medieval instance would not apply there.] As a parallel exercise, the brains of the chart-less indigenous mariners of today (discussed below) could be studied, with non-sailing members of their respective communities acting as a control. ^70a 

B.3b.1. London’s Black Cab drivers

The most obvious example of mental maps in action, albeit in an urban not marine context, is that of London’s Black Cab drivers. In what appears to those not involved to be an amazing feat, they create and then use a constantly updated, personalised mental map of 25,000 street names, 320 standard routes and 20,000 places of public interest, covering in all 20,000 miles [32,000 kilometres]. This process is aptly known as ‘The Knowledge’. Today’s drivers had learnt from an atlas, but they do not use GPS or carry paper maps. Their three- or four-year training is carried out when they are adults (usually supporting themselves with a job at the same time).

That needs to be contrasted with a medieval sailor’s focused experience, probably starting as a boy (when memories are more readily assimilated) and then spread out thereafter over perhaps half a century. Despite the fact that the majority of potential Black Cab drivers fail to complete the training, there are still 20,000 who have managed to pass and remain active today, keeping up with the steady stream of temporary or permanent changes in the London streetscape. In the light of that, it would be hard to contend that gaining a general (and in some parts detailed) mental picture of the Mediterranean would have been beyond the capability of a medieval navigator. ⁷¹ 

Initial research in 2000, and other studies since, have revealed that not only was the grey matter in the hippocampus of the cab drivers greater than that of London bus drivers (emphasising the importance of the cabbies’ decision-making) but that “the amount of time spent in the profession correlated with greater volume, proving the growth was accumulated”. This continued enlargement in brain function throughout the driver’s career reflects the fact that the stored information involved was both growing and changing. The memorisation process had involved transferring an existing street plan in its entirety into a mental construct. Conversely, the fact that people today increasingly rely on GPS to provide them with a precise location is apparently leading to a reduced volume in that same part of the brain. ⁷² 

The wayfinding skill and confidence of indigenous hunters who have adapted to their very different environments with varied navigation strategies, is at times extraordinary. Some of the Aboriginal ‘songlines’ – routes across the landscape linking features marking the tracks of the ‘creator-beings’ – are related to star positions. Where a Western astronomer might be able to name 100 stars, Aborigines, who had been steadily learning from childhood, can sometimes name thousands. When making long journeys across the Australian desert, the Aborigines were found to “possess some of the most precise orienting abilities known to man”. Their wayfinding skill involved “some kind of dynamic image or mental ‘map’, which was continually updated in terms of time, distance and bearing, and more radically realigned at each change of direction, so that the hunters remained at all times aware of the precise direction of their base and/objective” (reported by David Lewis). A person interviewed said he knew the directions “not by the sun but by the map inside my head”. One Aboriginal group in Queensland (and some other indigenous communities elsewhere in the world) do not have relative, subjective terms, such as left and right. Instead, they have sufficient confidence in their own spatial ability to use fixed, objective terms: North, East, etc. ⁷³ 

In Oceania, the piloting techniques used by various island populations in the vast spaces of the Pacific Ocean, are ably covered by Ben Finney in The History of Cartography (Volume 2, Book Three). ⁷⁴  In his introduction, on ‘Mental Cartography’, Finney summarises their attainments:

The navigational practices of Oceanians present somewhat of a puzzle to the student of the history of cartography. Here were superb navigators who sailed their canoes from island to island, spending days or sometimes many weeks out of sight of land, and who found their way without consulting any instruments or charts at sea. Instead, they carried in their head images of the spread of islands over the ocean and envisioned in the mind's eye the bearings from one to the other in terms of a conceptual compass whose points were typically delineated according to the rising and setting of key stars and constellations or the directions from which named winds blow. Within this mental framework of islands and bearings, to guide their canoes to destinations lying over the horizon these navigators applied vital information obtained by watching with the naked eye the stars, ocean swellsThe regular movement of rolling waves that do not break, steady winds, island-influenced cloud formations, land-nesting birds, fishing out at sea, and other cues provided by nature. ⁷⁵ 

For the Caroline Islands, Finney quotes from Thomas Gladwin:

Everything that really matters in the whole process goes on in his head or through his senses. All he can actually see or feel is the travel of the canoe through the water, the direction of the wind, and the direction of the stars. Everything else depends upon a cognitive map, a map which is both literally geographical and also logical. ⁷⁶ 

To Finney, the skills of the pioneering navigators, “must have included some competence in orienting and holding a course by reference to the stars, ocean swells, and winds, in dead reckoning, in sensing islands before they could be seen directly, and in incorporating newly found islands into some kind of cognitive chart”. ⁷⁷  To test those methods, “since 1976 voyages over 2000 miles of open sea have been achieved”. ⁷⁸ 

As another example, this time from the Canadian Arctic, Robert Rundstrom concluded that “the historical record and modern cartographic research both agree that most Inuit maps, extensively tested through a century of use by non-Inuit explorers and field scientists, were extraordinarily accurate renderings of the landscape as sensually perceived”. ⁷⁹ 

We need to be careful about over-stressing possible direct parallels between the experience of Mediterranean sailors and indigenous mariners. For example, the Cook Islanders used a 32-point wind compass, as does the portolan chart, though those are presumably unconnected. ⁸⁰  But neither should we deny the comparable demands made on all those who navigate over open seas seeking islands or distant shores. There must surely be some similarities in the solutions found and the mental capacity involved. There is no justification for assuming that the brains of indigenous navigators (whose skills have been observed up to recent times) worked in radically different ways to those of medieval Mediterranean sailors. It is always unwise to underestimate the ingenuity of those from a distant period, or a different culture, who either did not, or could not, leave a written record of their navigational practices.

One consideration that certainly links medieval navigation and indigenous wayfinding is that of necessity. In both instances, livelihood, safety, even life itself, depended on skill, memory, concentration, and careful observation when voyaging. In today’s world, “Navigation devices make vast reserves of distributed knowledge available to us in an instant. But, crucially, they never require us to possess information in our own memory in the way that successful navigators have been required to do till now”. ⁸¹ 

As one indigenous hunter put it: “Whereas most people like to go places, the Inuit people have to go places”. Another Inuit tracker, Taukie, gave an insight into the paired skills of memorisation and recall: “When we’re taking in details in our head, we try to look at the little stuff. When we travel past something we look behind us because it looks different from that angle. Every detail we are trying to put into our heads.” ⁸² 

As Barry Cunliffe pointed out, “It was 9000–5000 BC that saw the beginning of seafaring in Europe … The Phoenicians were the first of the Mediterranean States to enter the Atlantic in the 10th century BC and their main home port was Gadir on the Atlantic”. ⁸³  Another source locates the appearance of professional mariners and established routes in the Bronze Age,

the time where both archaeological records and texts confirm the beginning of the maritime specialization, that is, the appearance of sailors as distinctive professionals ... We can ascertain the existence at the IIIrd mill. BC, and especially for the IInd mill. BC, of established commercial waterways as it’s evidenced through the Ulu Burun and Cape Gelidonya wrecks. ⁸⁴ 

Thus, for centuries, indeed millennia, besides proceeding along a coastline (cabotage), ships had set sail across the Mediterranean (pelagic voyages), often out of sight of land, without the aid of navigational instruments, except perhaps a log line. ⁸⁵ 

Empires inexorably rise and fall, and maritime trade can be disrupted. But that was too vital for the parties concerned for it not to be continued, even if by indirect means. Hence there is no reason to suppose that seafaring practices – say those found effective by Roman pilots – would not have continued, possibly with improvement, up to the Middle Ages. Hence it is relevant to take note of what is known or conjectured about navigation in the ancient world.

We are fortunate that, with the highly literate classical world of Greece and Rome, we start to encounter written accounts, which are disappointingly missing for the medieval period. In his doctoral dissertation Danny Lee Davis made a thorough study of textual sources from the classical period. “In ancient literature”, as he explains in that major revisionist thesis:

“the oft-repeated themes of storm, shipwreck and death at sea led to the popular assumption among scholars that seafarers developed habits to minimize their exposure to this hostile element—hugging the shore to avoid the open sea, putting in at night, sailing only in summer, and using ‘seafaring manuals’ to help guide their way. While several recent studies have made some strides in overturning this overly simplistic view by highlighting aspects of navigation in certain areas and in certain periods, the ‘standard model’ lingers in both scholarly and popular imagination … My research concludes that both coastal and open-sea sailing were matters of routine in the commercial sector, that commercial seafarers did indeed sail at night and employ the stars to deduce navigational information, that winter sailing was a widespread practice, and that crews employed navigational strategies to weather storms, usually successfully.” ⁸⁶ 

When writing about navigation in the later Middle Ages, Renard Gluzman confirmed Davis’s account, noting that “crossing open water was an integral part of any long-distance voyage … Venetian vessels sailed greater distances and much farther from the coast than had previously been thought”. ⁸⁷ 

Instead of enquiring whether the Greeks and Romans had marine charts, we should rather ask, did they produce graphic versions of the charts they apparently held in their heads, to which the answer is that there is no evidence of that.

If landmarks were not going to be available as a guide, in the period before the portolan chart, pilots had to know beforehand, fairly precisely, in which direction(s) to head and roughly how long it would take to reach their destination. Another obvious point, which might be overlooked, is that for land to be recognised it had to have been anticipated. When planning a voyage, the pilot would need to take account of intervening islands, note where the voyage might be broken, or water and supplies shipped, and in particular, anticipate the dangers that were likely to be encountered en route. Any lesser assumption is surely untenable. Unless they had followed memorised courses, no routine trading, pilgrimage or military voyages could have taken place. ⁹⁰  Indeed, these ‘sea roads’ are well documented from surviving texts and wrecks (which were more likely to have been caused by weather conditions than poor seamanship). The following pair of assertions that underpins this essay is therefore proposed:

The mental abilities required for medieval navigation would fall into those two types described earlier in the discussion on mental maps: respectively, route knowledge and landmark orientation. ^90a Because of the risks involved in open-sea journeys, navigators needed to acquire a packed memory of courses run and dangers avoided, backed up by the visual appearance of the numerous headlands, bays, river estuaries, coastal and island outlines, etc., they had seen. William Bourne, in his Regiment of the Sea (1574), noted that a shipmaster should “be a good coaster, that is to say … knowe every place by the sight thereof”. ⁹¹ 

Cunliffe wrote that:

For any sea or stretch of coast, the accumulated sea lore would have been vast. It was essential to the well-being of maritime communities that it was kept up to date and passed on from one generation to the next. In preliterate times this would have been done by word of mouth, possibly in the form of chants or songs, or at least in rhythms that allowed the instructing to be committed to memory. ⁹² 

The record of each newly completed voyage would need to be memorised, to join those from previous courses run, since remembering those details would be crucial when any came to be repeated.

Danny Lee Davis makes a number of interesting observations on routes actually taken as compared to those theoretically described, a distinction not generally recognised by historians:

Sailing vessels rarely traveled in straight lines. Instead of fixed routes sailing masters chose to sail within seasonal corridors of movement in which conditions were most favorable for efficient forward movement. ⁹³ 

He also pointed out that “No route, no matter how short or how often made, was ever repeated precisely. At odds with this notion, however, are the ubiquitous maps of finely delineated sea routes we find in studies of ancient seafaring, trade and economics”. ⁹⁴  In other words, the focus on simplistic, generalised descriptions of pelagic routes may be misleading. Though his remarks derive from a detailed examination of Classical navigation (well supplied with textual sources), his elaboration of the argument is worth repeating for its likely application to the largely undescribed medieval context.

Based on our source material, and taking into account the multitude of variables involved in each voyage even in optimal conditions, we should instead envision ancient sea ‘routes,’ whether short- or long-haul, as wide maritime corridors of general movement between one place and another. These corridors were defined by environment and meteorological factors, by technological responses to the demands of sea travel and by the ever-shifting realignments of trade trajectories throughout the seasons, years and centuries. The aggregate effect was of loose bundles of overlapping lines connecting coastal and island nodes, and not of straight, single paths overlaced multiple times. ⁹⁵ 

C.2b. The skill-set required for navigation

We cannot hope to understand how medieval sailors navigated with a chart without knowing how they had managed without one. In other words, what did portolan charts offer pilots that they didn’t already have?

Prior to the chart and the magnetic compass, medieval pilots would have needed to interpret the sensory information they were receiving so as to answer three basic questions: where are we, where do we want to go, and how will we best get there? Navigating with a wind compass ⁹⁶ might have been equivalent to a blind commuter today finding their way to work by means of remembered features and sensual clues, not available to their sighted colleagues. The focus needs to be more on those aspects, undescribed and almost entirely lost, rather than the well-documented, tangible mechanical aids.

Davis offers a relevant parallel from the Classical world:

Historical sources reveal an occupation of low social status that nonetheless required a prodigious set of complex skills and a practical knowledge earned from numerous hard years spent at sea—seamanship, crew leadership, maritime geography, winds, currents, weather prognostication and nautical astronomy, among others. Together these skills and knowledge constituted the “steersman’s art,” ta kybernetika, or the ars gubernatoris. ⁹⁷ 

When establishing the approximate position at sea, the 12th-century mariner would have used ‘dead reckoning’: the ability by long experience to estimate the direction and distance covered, by relating the ship’s apparent speed through the water to the elapsed time, with the use of astronomical and other indications. He would have had no instruments, apart from a log line. It seems likely that the magnetic compass was developed after the time of the first portolan charts ^97a and the hourglass or sand clock somewhat later still. ^97b  To determine the ship’s position, the calculations of distance and direction (using such things as the angle of the swellThe regular movement of rolling waves that do not break) were augmented by astronomical aids, and particularly the known bearing of the expected wind (which had provided sailors with a theoretical compass for centuries before they first saw a magnetic needle).

When estimating the ship’s speed, allowance would be made for a following or adverse wind, the swell, any current there might be, and the effect of tacking. As far as direction was concerned the mariner had to select the desired heading, maintain it, and, if forced off course, regain it. For this, the toleta de marteloio, ‘a trigonometric reduction table to aid navigation by dead reckoning’, ⁹⁸ would have been relevant but it is not clear how widely it was used to calculate the actual course sailed when tacking into the wind. The oldest known textual reference to it is from 1295, though that does not mean it could not have been in existence earlier. ⁹⁹ 

C.2c. Sea signs

When cloudy conditions obscured the sun or the stars, navigators would have had to rely on their own senses. Even if, when calculating distance and determining direction, they may not have had the equivalent of perfect pitch in music, it must surely have been the case that those senses would have been heightened first in training and then through constant use. Medieval pilots would have been attuned to their surroundings far more acutely than those of us in the developed world today, even experienced mariners. The sophisticated layers of personal and very detailed information, which must often have been retained in visual form, would also have included a wide variety of sea sign. The following would be a probable, if incomplete, list of those:

Approaching land: signs such as orographic clouds (created by the topography of an island that might still be out of vision); sightings of birds, fish or aquatic mammals; ¹⁰⁰ driftwood, seaweed or floating vegetation.

Coastal profiles: fixing location in relation to the appearance of a coastline would have depended on reference to what must have been a large part of a navigator’s visual memory

Currents: the portolan charts could have introduced a symbol for areas with strong, regular currents, for example between some islands and the mainland, but they did not. So there is no warning on the charts, for example, about the strong tidal currents in the Strait of Messina dividing Italy from Sicily with its Charybdis whirlpool, nor the Gibraltar and Dardanelles straits. The implications of those would have to be learned but, once experienced, would surely not have been forgotten

Depths: checked by soundings with a lead line, perhaps revealing the bottom deposit which might help to identify the area; measured soundings were only introduced to charts in later centuries, but the charts’ hazard symbols served a related purpose by highlighting dangerous shallows

The realisation that the details in the Liber’s pelagic statements must have been obtained from a prototype chart of some kind was a major discovery. ²⁶⁴  Each of the single directions, followed by a mileage figure – neither of which would have reflected the realities of a long voyage – could have been obtained only by reading the bearing off the wind compass network of a marine chart and scaling the distance from the same source. But that on its own does not wholly answer the ‘chicken and egg’ question about the primacy of text or chart. Even more convincing is the parallel evidence from the portolani (Singular portolano and more usually called ‘portolans’ – which can be confused with the charts) these were medieval guides to the coastal places around the Mediterranean, listing the distances between them. They also include open-sea (pelagic) routes’s coastal itineraries.

The debate as to the priority of the portolani or the portolan chart is clearly in favour of the latter. Joaquim Alves Gaspar’s assertion, quoted above, ²⁶⁵ is fully vindicated. The stated directions in the LiberThe ’Liber de existencia riveriarum’, c.1210, written in Latin, is the oldest surviving guide to the Mediterranean, giving the distance and direction between places along the coast and across the open sea (Cotton MS Domitian A XIII, ff. 114r–129v) and Lo compasso’Lo compasso de navegare’, c.1260-96, written in Italian and apparently designed for sailors, is much more detailed than the earlier ’Liber’ (Staatsbibliothek, Berlin, MS Hamilton 396) must have been straight lines calculated by eye from a chart’s network of compass lines [Powerpoint, Slide 15], and the distances would have been measured with a ruler. In other words, not only must a marine chart have preceded the oldest portolano that included direction and distance statements, but, since the chart was the source for those, it must have partially enabled the development of such texts.

For uncountable centuries before the advent of the portolan chart, sailors must have gone through a similar, but very personal, never-ending learning process during which they constructed an approximate picture of the Mediterranean as a whole, and the specific placement of its many islands within that. No pelagic sailor who had responsibility for navigation could have done otherwise. And no landsman – at least judging by the maps that survive from before 1200 – could have had access to outlines remotely equivalent to those memorised by pilots. Compiled in very different ways, al-Idrīsī’s Charta Rogeriana ²⁶⁶ and the maps based on Ptolemy’s Geographia demonstrate the constraints facing those viewing from the land when trying to plot the circumference of that sea (or, to put it another way, the continents’ landward limits).

One of the challenges facing a cartographic historian is to see the world through the eyes of the past. For us, the question ‘what does the Mediterranean really look like’ is redundant. We start with a single possible answer, particularly after our planet was photographed from space. It is almost impossible for older people to forget that image, and remember how previously we had to rely on maps taken on trust. But how would a savant in 1200 have been able to adjudicate between, say, Ptolemy’s maps (had they been available), the mappaemundi’World maps’ in Latin (singular, mappamundi). Used to distinguish medieval maps of the world from those of Ptolemy and the portolan charts or the Islamic depictions? Which was better? Which was ‘right’? ²⁶⁷  The navigator, on the other hand, with his systematically gathered and carefully stored memories, regularly reinforced by further voyages, was in the unique position of being able to answer those questions with confidence. Not, of course, because he had ‘seen’ the Mediterranean in its entirety but because he had mapped out or learned much of it, piece by piece, and stored the results in his head.

It is worth going back to basics and asking the question: how might a medieval mariner have envisaged the Mediterranean? The sailors’ viewpoint is both horizontal (across the waves) and vertical (when examining in profile an island they were passing or the coastline appearing over the horizon). By contrast, a portolan chart is, in effect, the coastal outline extracted from an aerial view: in other words what we take for granted in a planimetric map. The charts of Lucas Jansz. Waghenaer, in his printed Spieghel der Zeevaerdt (1584) combined two of those viewpoints, with rough profiles of cliffs placed beside the coastal outline.

For a traveller on land, the roads and rivers provide a horizontal network that can be readily imagined as a map, but there was no equivalent aid for a sailor. How would the charts’ creators have been able to translate what they were seeing into cartographic form, particularly, as is suggested elsewhere, if they might never have seen a map of any kind before? ³⁰⁶ 

The way that the putative mariner(s) who devised the portolan chart (probably at the end of the 12th century) arrived at that aerial view, extrapolated from the profiles they had observed, is one of their major achievements. Even in the 15th century, as Matthew Boyd Goldie pointed out: “An unlimited and an overhead aerial perspective and understanding of space was possible for people at this time, but it was not how people customarily thought about or presented the areas around them”. ³⁰⁷ 

Underlying this essay is the thesis that portolan charts owe their structure to the memorised pelagic courses, which had placed the features of navigational significance in the Mediterranean and Black seas in realistic relationship to one another. From that, it follows that the charts are projectionless, as others have previously suggested. ³⁰⁸  What intermediate mathematical calculations could have intervened between mariners’ memories and the graphic record of those? ³⁰⁹ 

Joaquim Alves Gaspar elaborates further on the portolan charts’ lack of a formal projection:

Specifically, it made it possible to demonstrate that the main geometric features of portolan charts are explained by assuming that the navigational data—compass courses and estimated distances—used in their construction were transferred directly onto the plane of the chart irrespective of the fact that the measurements had originally been made on the earth’s spherical surface. While no medieval text confirms this hypothesis, early modern texts explicitly refer to charts of the Mediterranean being constructed by using compass courses and estimated distances collected by pilots. ³¹⁰ 

On the question of the portolan charts’ projection, or lack of one, Roel Nicolai’s is a dissenting voice. The issue is much discussed in his 2014 thesis and was summed up thus at the end:

The map projection found in portolan charts cannot be the unintentional result of a simple chart construction method or an accidental by-product of any cartometric analysis process. The follow-on conclusion is therefore that, in the absence of any convincing alternative explanation, the map projection is not accidental, but must have been intentionally designed into the charts as part of their construction. ³¹¹ 

That needs to be considered alongside another statement by Nicolai that: “The construction of a chart with an a priori intended map projection would definitely be well beyond the capabilities of medieval sailors, cartographers and even the intellectual elite of that period”. ³¹²  Nicolai’s repudiation of the possibility of a medieval origin for the portolan charts has been challenged throughout this essay.

This brings us close to the crux of the portolan chart enigma: the missing link between, on the one hand, an imagined prehistory containing a number of logical but largely uncorroborated developmental phases and, on the other, by the oldest surviving example of the genre, as well as two texts, each unique, each of uncertain dating, and each open to differing interpretations. Additionally, the novelty visible in a series of extraordinary conceptual leaps, which might have involved more than one person at more than one time, and the total absence of the expected ties to a previous cartographic tradition, means the range of options is a large one.

Here I immediately diverge from some other commentators. Whereas it is argued here that pelagic memory was the ultimate source, Ramon Pujades (and others) argue that the charts had a textual paternity. In his contention he refers to “data transposed from the written to the graphic format, which constituted a genuine revolution from the intellectual viewpoint”, which he refers to elsewhere as “the major exercise in mental abstraction involved in the transposition of a written compilation of data on distance and direction to the graphic format of a scale map endowed with a wind network”. ³¹³ 

Pujades makes a specific conjecture that it was the clergy who played this crucial linking role, since as a sector they were “directly involved in the task of compiling and organising information which, amassed by sailors from their experiences at sea, would initially have been restricted to the oral realm”. He goes on to suggest that “in all probability, the author of the LiberThe ’Liber de existencia riveriarum’, c.1210, written in Latin, is the oldest surviving guide to the Mediterranean, giving the distance and direction between places along the coast and across the open sea (Cotton MS Domitian A XIII, ff. 114r–129v) was one of the numerous clerics who acted as notaries and chaplains on the naval convoys”, and he finally makes a specific proposal that “the first navigational chart makers would have belonged to this same socio-professional group.” ³¹⁴ 

Nevertheless, no fragment of any text or diagram has survived, or at least been recognised as such, neither has a working record of a calculation using a toleta de marteloio‘A trigonometric reduction table to aid navigation by dead reckoning’, used to calculate the actual course sailed when tacking into the wind. Proposed theoretically by Gregory McIntosh (at the 2016 Lisbon workshop) this would have been “a piece of parchment for drawing the sides of triangles for recovering the course” when tacking.

Expressing a strictly artisanal view in a recent article, Joaquim Alves Gaspar and Henrique Leitão observe:

Pre-Mercator nautical charts are complex objects whose inner geometrical properties are only now starting to be understood. Rather than attempts at a faithful geographical depiction of the world, they were tools aimed to accomplish a specific purpose – to support marine navigation – a fact that is often mentioned in contemporary sources. Thus, they carry inside their internal geometry the imprint of the activities of pilots on board, that is, a ‘signature’ of their artisanal origin and of the techniques used to navigate ... nautical charts were deeply influenced by the practices of the artisans who used them, but, simultaneously, such influence was almost invisible to contemporary scholars and present-day historians … first, that historically, nautical charts can only be understood in the context of the specific navigational methods they were intended to support; and second, that a nautical chart should not be considered as a true geographical map but as an instrument for navigation. ³¹⁵ 

Support for the notion that the missing link might have been a diagram rather than a marine chart, can be found in the Liber’s puzzling descriptions of the Tuscan island of Pianosa as quadrangular and Cyprus as a double quadrangle; those terms are more suggestive of classical textual sources. Similar instances are the triangular Sicily and square Cyprus found in some mappaemundi ’World maps’ in Latin (singular, mappamundi). Used to distinguish medieval maps of the world from those of Ptolemy and the portolan charts. ³¹⁶  Those would not have been helpful characterisations for a sailor. ³¹⁷  But descriptions of that kind would make more sense if somebody was looking at a simple pelagic diagram emphasising the capes that defined an island’s overall shape, but before any attempt had been made to delineate the coastlines.

Even if the hypothesis that navigators had been able to build up a comprehensive mental map could be proved, that would, on its own, have little relevance for the origin question. Not until the information had been ‘downloaded’ into graphic form could it have served its conjectured purpose as the foundation for a simple cartographic diagram that could develop into a marine chart.

We might conjure up the germ of the portolan chart in a fanciful exchange between an enquiring passenger (perhaps a pilgrim) and an experienced navigator, in an incident akin (though reversed) to that of an explorer asking an indigenous guide to sketch out a map on the ground. In this instance the questioner might have asked: “Please tell me how far we have come, what places we will be going to, and how long you think it will take?” Drawing out the course, assuming the necessary materials were available, would have surely been the easiest way to answer those questions. As a possibly comparable example, the 10th-century member of the Balkhi School, al-Muqaddasī described how an Arab merchant in Aden, when he was asked for information on the Indian Ocean, “rubbed the sand with his palm and drew the sea on it”. Although the image was ephemeral, it did demonstrate the process of downloading a mental chart. ³¹⁸ 

If such a pelagic diagram existed it would have been logical for it to have been sketched over a circular wind network, where each of the eight or sixteen wind directions ran through the centre of the circle. If so, that would prefigure the network which underlies all surviving portolan charts. ³¹⁹  It would also mean that the same directional pattern that sailors would later use to lay off their course had previously helped to create those charts in the first place.

How such a diagram might have been drawn is a matter of pure speculation. Vellum was an expensive material (hence its routine reuse as a palimpsest) and it is unlikely that a prepared skin would have been employed for a preparatory pelagic diagram. Nor would paper have been readily available to European sailors around 1200. ³²⁰  al-Muqaddasī again provides an interesting 10th-century parallel, when he reported seeing two maps of the Indian Ocean, drawn respectively on paper and cloth. ³²¹ 

A more plausible material on which to lay down a geometric diagram of the Mediterranean might have been some kind of textile although the coarseness of sailcloth could mean we have to look elsewhere for the most likely medium, perhaps discarded clothing. On the basis that the diagram would have been constructed from plotted-out courses drawn or described by mariners, it could have been created just as well at sea as in a port. Sailors, whether literate or not, are unlikely to have had access to a quill pen and ink but among the crew there might well have been clerks or others who could have provided those. ³²² 

The main, and most challenging question has always been to provide an even semi-plausible explanation as to how the components for a complete and reliable chart managed to be thought out and laid down cartographically. As a partial recapitulation of what has already been said above, the process of transferring information held in memory into some kind of graphic form, and then elaborating on that, is likely to have had a number of distinct stages, each representing a significant conceptual leap:

1. It might have started with a single long-distance course, though that could have taken one of two forms: either a simple straight line (perhaps conveyed orally), or a more elaborate record of the stages of an actual journey.
2. That would need to have been followed by the addition of further courses sharing the same pelagic terminus, which would have allowed the possibility, in time, for the design to grow, incrementally, into an extended network. At that point those involved would have presumably noticed that some repeated termini had been located in slightly different positions. This would have alerted them for the first time to the existence of those errors, thus allowing them to be corrected, whether by moving the terminus along the coast or by adjusting the pelagic distance. It is open to speculation as to how much of that hypothetical network of pelagic courses could have been constructed mentally and how much would have become apparent only once it was in graphic form.
3. At that stage there might have been the crucial insight that, by setting down those details on a flat surface, the overall pattern would become clear (along with the lacunae) and new routes could be extrapolated from the existing data. Thus a diagram created from memorised pelagic courses could later be used to plot future ones.
4. Presumably that would have triggered the realisation that coastal outlines could now be sketched in between the already-placed headlands.
5. That, in turn, would have allowed the addition of the sequence of place-names, a large proportion of which would presumably have formed part of the sailors’ mental store anyway. The imprecise positioning of the toponyms on a portolan chart might endorse the contention here that the names had merely been memorised in the form of a sequential list rather than, as in a portolano (Singular portolano and more usually called ‘portolans’ – which can be confused with the charts) these were medieval guides to the coastal places around the Mediterranean, listing the distances between them. They also include open- sea (pelagic) routes, being placed in relation to the size of the intervening distances between them.
6. To be fully functional, the chart also needed to be furnished with a range of conventions, most of which were then apparently unknown to cartography. Among the most obvious of these are an underlying network of 32 wind or compass directions; an indicated map scale; the absence of colouring for the seas; toponyms consistently placed at right-angles to the coast (the more prominent of those written in red), giving the chart a flexible orientation that matched the way it would be turned during use at sea; and the beginnings of a coded symbology for navigational dangers. ³²³ 
7. In addition to the post-Carte PisaneAccepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) adjustment of the Adriatic, and refinement to the hazard symbols, the other feature which still required development was the handling of the spandrels on the twin-circle networks (the area above and below the point of tangency). This was left un-provided with compass lines on the earliest charts. The Carte Pisane and Avignon chartA chart fragment focusing on the North Sea, probably very early 14th century, discovered in 2002 in the Vaucluse Archives (Port 01, 3E 54) – for Jacques Mille’s studies on this see the Portolan Chart Bibliography: https://www.maphistory.info/portolanref.html chart filled in that space awkwardly with supplementary grids, while the Cortona chart’sBased on its archaic appearance, this is considered to be the second oldest surviving portolan chart after the Carte Pisane, dating from the beginning of the 14th century or even earlier. It was first described in 1957 (Biblioteca dell'Accademia Etrusca di Cortona, Port. 105) network ignored the Adriatic entirely. Later charts experimented with a variety of ingenious geometric arrangements for the junction of the twin circles, demonstrating both the still relatively undeveloped nature of the Carte Pisane and the continuing creativity of those unknown early practitioners.
8. Finally it must have been appreciated that an animal skin, largely untrimmed, might be used to carry a reproducible map of the Mediterranean. [Is there any evidence that entire, unevenly-edged skins had been used by scribes before?] The holes in the neck of the Carte Pisane may indicate a further reason for not trimming the skin down to a neat rectangle. Once it was rolled up, it could then be secured with a strip of leather anchored in that hole.

When the chart first came to be drawn out it is likely that east or south would have been prioritised in its orientation, since dawn and the midday position-sighting are likely to have been the main daytime events for the navigator. However, the shape of the Mediterranean invites a ‘landscape’ presentation (coincidentally matching the animal skin on which the completed charts would later be drawn). ³²⁴  There are some indications that, when choosing whether to put north or south at the top, the early charts favoured a southward orientation.

This transition from a privately-stored mental network of pelagic courses to a drawn-out diagram is crucial to the thesis of portolan-chart origin set out in these pages. “Like any technical object”, Patrick Gautier Dalché observed, “the marine chart represents an object de pensée (a materialized thought-object)”. ³²⁵  As we are contending, it would not logically have been possible for the portolan chart to have developed from any other source than the memorised knowledge of pilots – at least before, say, the 16th century, when adequate navigational instruments had become available. ³²⁶  Such a metamorphosis from mental construct to graphic reality can also best explain the “sudden emergence of nautical cartography”, remarked upon by Ramon Pujades. ³²⁷ 

Until the portolan chart (or its antecedents) emerged, maritime lore must have been broadly static, equally likely to deteriorate as to improve, with the hard-won spatial knowledge of an individual mariner disappearing at his death. Instead, with a graphic alternative, a shared knowledge could be enriched by subsequent generations, in a way similar to that experienced in a literate environment. The change from virtual to tangible allowed information to be first shared, then refined, and subsequently augmented. In that way, the medieval pilots were able to break free from the need for the constant re-invention of the mental map. Once provided with a portolan chart, a ship’s boy might have started with considerably more information about the hydrography of the Mediterranean than a 50-year-old mariner would previously have had locked in his head.

I.4c. The rectification procedure

A single remembered course could never have been more than approximate but, once combined with others originating or concluding at the same points, errors could have been steadily cancelled out. Lo compasso’s’Lo compasso de navegare’, c.1260-96, written in Italian and apparently designed for sailors, is much more detailed than the earlier ’Liber’ (Staatsbibliothek, Berlin, MS Hamilton 396) pelagic statements include a number of multiple instances of such duplication. About twenty instances, for example, start or end at Malta [or Gozo] ³²⁸ as can be seen graphically in Roel Nicolai’s diagram. ³²⁹  When faced with conflicting placements of the pelagic termini involved in multiple incoming and outgoing transits, a graphical optimization process could have averaged them out. ^329a  Subsequently, the common position could have been tested on later voyages and, where necessary, rectified further. Hence the reliability of the proposed pelagic diagram would not have depended entirely on the ability of medieval mariners to measure distance precisely.

This hypothetical process of positional refinement, although not yet generally accepted, has been stated before, in this case by Ramon Pujades: “It would have been necessary to compile the information culled from a substantial number of crossings, because perfection in terms of representation could increase only as the number of points grew that could be located in respect to each other with relative precision”. ³³⁰  Since each pelagic course is in fact two (outward and inward), this would have doubled the possibilities for comparison with an earlier estimate, and as a result lead to possible rectification – part of the necessary process of collective endorsement.

Some of those revisions could have already taken place mentally, although there must have been limits to the complexity of what pilots would have been able to retain in that form. Most would probably have been subsequent additions and refinements, which could have been made only once there was an initial graphic diagram on which to record them. Writing in 1279, near the time that the Carte PisaneAccepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) was probably drawn, Egidio Colonna commended the realistic geometry of the portolan chart, writing that sailors “draw a map of the sea on which sea ports, hazards and other such things are described in proportion”. ³³¹ 

Because of the lack of intervisibility it would have been natural for there to have been separate mental maps for the different Mediterranean basins: in other words those that were discrete entities in terms of sailing. ³³²  There would no doubt have been gaps and many inconsistencies – but that could actually have provided part of the impetus for a shared effort to record the network, complete it, and then iron out the mismatched positions. ³³³ 

While the Nicolai thesis (logically if not explicitly) denies the possibility that the medieval mariner could have improved what he was unable to create, we can actually see that process of rectification taking place on the charts that followed the Carte Pisane – most notably in the width of Italy and the Adriatic. ³³⁴  Although the broadened shape of Italy is what catches the eye, the error actually concerned the relative position of the unconnected coastlines that frame it. In other words, it was not the result of faulty observation from the seas either side of Italy, nor could the error have been readily appreciated or amended by navigators.

Long after the development stage, in 1403, Francesco BeccariFrancesco and his son Battista were active in or around Genoa in the first half of the 15th century. Francesco’s 1403 chart is notably innovative (New Haven, Yale University, Beinecke Library) was to explain that he had been informed about the incorrect positioning of Sardinia. ³³⁵  The way in which that error (if it was indeed one) was identified by informants is likely to have mirrored the earlier rectification process when mariners’ experience must have challenged the positions of pelagic termini on the emerging chart.

Moving on to the replication stage, we might assume that the charts would initially have been copied by the same individuals who were involved in developing them. As the demand grew, a stage must have been reached when the nascent charts were considered sufficiently accurate and complete to be worth wider circulation. The fine and varied detail required by the charts called on special drafting and scribal skills for what was essentially several different documents in one: an underlying mathematical structure formed of compass lines; coastal outlines; navigation danger markings and written toponymy, requiring experience at using inks of different colours; and, as time went on, colour washes. ^335a  To preserve the charts’ main purpose, namely its geometric accuracy, only strict workshop discipline could have avoided progressive distortion. This was the main key to the charts’ longevity.

Among the charts produced in 15th-century Venice were some whose creators were also sailors, but it is unlikely that a mariner from two centuries earlier would have possessed the required scribal skills. Pietro Vesconte (fl.1311-c.30) is the first chartmaker known to us by name, and the first professional. A Genoese, he worked in Venice, along with his presumed relative Perrino. Several of his works survive, documenting in particular the growing shape and toponymy of the British Isles. His chart of 1311 and atlas of 1313 are the oldest to be signed and dated displays a neat and very confident hand, ³³⁶ and he had clearly been trained as a scribe. However, there is no indication that he had served at sea; indeed, that seems improbable.

• Sailors, probably unlettered, gathering, over time, a carefully memorised knowledge of the pelagic courses and, no doubt, of the broad outlines of the coasts and the toponyms ranged along them
• Imaginative mariners who were able to transform those mental maps into a graphic form and thus, again over time, draw out a coherent chart of the entire Mediterranean and Black seas
• Trained copyists who would have been required once the demand for charts created a viable market for the replicas
• Genuine ‘chartmakers’, i.e., those who were in control of the charts’ content and who could, and perhaps did, make alterations and additions. We know nothing about the earliest practitioners until Pietro Vesconte – whose work shows steady development – signed his chart of 1311, some decades after the Carte PisaneAccepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118)
• Apprenticeship (often, no doubt, with a family member) must have included a thorough training in the house-style used for all aspects of chart construction and particularly the toponymic lettering
• Since the chartmakers were port-bound, they must have been dependent on anonymous informants (whether casually or by invitation) for fresh details with which to update the charts, particularly their toponymy ³³⁷ 

Coverage of the Black Sea in Ptolemy’s Geographia is split between several sheets: Prima Asiae (southern coast), Prima & Tertia Asiae (east coast), Secunda Asiae & Octava Europe (Crimea and the north coast), and Nona Europe (west coast). ³⁴¹  Unless the sections are joined up, this means that only the world map provides the whole sea in a single view, necessarily at a small scale. Overall, Ptolemy’s version is surprisingly plausible, and certainly superior to the earliest attempts by the portolan-chartmakers, avoiding the distorted shape of the south-east section, as described below. This supplies evidence that even if Ptolemy’s work was being used it was not yet being imitated. ³⁴² 

Like some other Islamic geographers, al-Idrīsī is thought to have made use of the Ptolemaic outlines, but that could not have applied to the Black Sea. The unrecognisable shape he gives to that sea, with its imaginary islands, some treated as equivalent in size to Corsica, could not have borrowed anything from Ptolemy.

J.1b. 13th-century developments

It might have been hoped that 13th-century records would have helped to reveal how the Black Sea’s outlines were captured. Unfortunately, the available sources fall between the slender and the non-existent, and what there is comes without secure dating. There is nothing between the LiberThe ’Liber de existencia riveriarum’, c.1210, written in Latin, is the oldest surviving guide to the Mediterranean, giving the distance and direction between places along the coast and across the open sea (Cotton MS Domitian A XIII, ff. 114r–129v) , with its meagre 13 toponyms, and a more detailed pilot guide, Lo compasso’Lo compasso de navegare’, c.1260-96, written in Italian and apparently designed for sailors, is much more detailed than the earlier ’Liber’ (Staatsbibliothek, Berlin, MS Hamilton 396) with 104. Besides those two witnesses are a small group of early, unsigned and undated, charts which might date to the period between the Carte PisaneAccepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) and the first dated chart, Pietro Vesconte’sPietro Vesconte (fl.1311-c.30) is the first chartmaker known to us by name, and the first professional. A Genoese, he worked in Venice, along with his presumed relative Perrino. Several of his works survive, documenting in particular the growing shape and toponymy of the British Isles. His chart of 1311 and atlas of 1313 are the oldest to be signed and dated of 1311. ^342a  Whereas the process of defining the Atlantic coasts of the North Sea can be progressively documented in charts that followed the Carte Pisane, in the Black Sea instance. because of a lack of written evidence, we have no witnesses to the early years of its surveying history.

Lo compasso, whose text (incorporating amendments) is thought to date within the period 1260-1296, includes a section on the Black Sea. ³⁴³  However, that comes at the end of the volume and in Alessandra Debanne’s opinion was ‘probably’ missing from the original form of that work, only being added for the surviving example. ³⁴⁴  Since the volume bears the date 1296, it is hard to argue that the Black Sea section is any earlier than that, thus placing it, in this context, later than the Carte Pisane. On that basis, this leaves the period 1210-70 without visual or textual evidence about the cartography of the Black Sea. ³⁴⁵  Investigation is further hampered by the Carte Pisane’s twin limitations: that its eastern third was (intentionally) omitted, and that serious damage to the vellum in the Black Sea area means that some outlines and a number of place-names have been lost.

J.1c. Toponymy

The resulting omission from the Carte Pisane of Tana, Savastopoli, Trapesonda, etc., might seem to imply a lack of awareness of their significance. However it is more likely that the author, with restricted space available, kept to the dimensions of earlier charts, whose eastern edge followed the longitude of the Levant coastline, while making sure, at the western end, to include all of the Iberian Peninsula. ³⁴⁶  As a result, it is no longer possible to get a sense of the sea’s overall shape on the Carte Pisane, and much of our knowledge of its toponymy derives from tentative readings on a facsimile made in 1852, before some of the damage had occurred. ³⁴⁷ 

Toponyms are easier to compare than coastal outlines but in this case they are an imprecise guide to the cartographic development of the Black Sea. No more than 25 Black Sea names can now be retrieved from the Carte Pisane. Yet, of that small number, ten do not appear on Lo compasso and eight are absent from the (supposedly) second oldest survivor, the Cortona chart,Based on its archaic appearance, this is considered to be the second oldest surviving portolan chart after the Carte Pisane, dating from the beginning of the 14th century or even earlier. It was first described in 1957 (Biblioteca dell'Accademia Etrusca di Cortona, Port. 105) while seven are not repeated on either. Lo compasso and the Cortona chart have an identical number of Black Sea toponyms (104), which might point to conformity. However, in each case half of their total is not found on the other. In other words, there was a marked lack of uniformity in the second half of the 13th century about the selection of names around the perimeter of the Black Sea. As happened elsewhere, the Vesconte chart of 1311 and his atlas of 1313 mark the beginning of what would become a stable toponymic sequence, increasing steadily afterwards. ³⁴⁸ 

The predictable way to have surveyed the Black Sea would have been as a two-pronged effort, by documenting the coastal locations and their intervening distances, on the one hand, and building up a geometric network of pelagic courses so as to fix the sea’s perimeter, on the other. Lo compasso ’Lo compasso de navegare’, c.1260-96, written in Italian and apparently designed for sailors, is much more detailed than the earlier ’Liber’ (Staatsbibliothek, Berlin, MS Hamilton 396) , which includes over 1,200 open-sea pelagic courses (Peleio) for the Mediterranean, omits those entirely for the Black Sea. As, indeed, do later portolani (Singular portolano and more usually called ‘portolans’ – which can be confused with the charts) these were medieval guides to the coastal places around the Mediterranean, listing the distances between them. They also include open-sea (pelagic) routes, which further argues against there having been any systematic documentation of pelagic tracks in that sea. Since the coastal outlines on the Carte PisaneAccepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) are now largely illegible, this means that the first available marine chart of the Black Sea is represented by the Cortona chartBased on its archaic appearance, this is considered to be the second oldest surviving portolan chart after the Carte Pisane, dating from the beginning of the 14th century or even earlier. It was first described in 1957 (Biblioteca dell'Accademia Etrusca di Cortona, Port. 105).

The lack of geometric underpinning is confirmed by distortions to the overall shape on the Cortona chart, particularly in the eastern half, which is noticeably misaligned. The other undated charts that followed the Carte Pisane are not relevant: the Avignon chartA chart fragment focusing on the North Sea, probably very early 14th century, discovered in 2002 in the Vaucluse Archives (Port 01, 3E 54) – for Jacques Mille’s studies on this see the Portolan Chart Bibliography: https://www.maphistory.info/portolanref.html fragment does not extend east beyond Italy, and the Lucca chartDiscovered in 2000 in the city archives, this is visually very different from other charts from the 1st quarter of the 14th century (Archivio di Stato, Lucca, Fragmenta Codicum, Sala 40, Cornice 194/1) includes only the south-west portion of the Black Sea.

Hence the focus must turn to the 1311 Pietro VescontePietro Vesconte (fl.1311-c.30) is the first chartmaker known to us by name, and the first professional. A Genoese, he worked in Venice, along with his presumed relative Perrino. Several of his works survive, documenting in particular the growing shape and toponymy of the British Isles. His chart of 1311 and atlas of 1313 are the oldest to be signed and dated chart and that in the Riccardiana LibraryThe oldest Genoese chart, probably dating from the second decade of the 14th century (Biblioteca Riccardiana, Florence, 3827), perhaps of roughly the same date. Those show similar improvements to the shapes of Crimea, the Sea of Azov and the south-east, but there is still no appreciation that the lowest points along the south coast (near to either end) are on the same latitude, or that the moving up of that coastline has caused the east coast to be compressed. Nevertheless, that imperfect outline must have been thought acceptable, since it was left uncorrected and can still be seen in the work of Grazioso Benincasa An ex-ship captain, he left us a range of clearly-drawn charts and atlases (1461-82) and a written portolano (1435-45) in the 1460s, and on subsequent charts. As a result, after being worked on for up to a century – with a magnetic compass certainly available for much of that time and a relatively uncomplicated coastline – the outline of the Black Sea ended up as less realistic than those that had been achieved across the entire Mediterranean (and particularly the Adriatic and Aegean seas) prior to 1204.

The likely explanation for what might not perhaps have been considered as a shortcoming by the charts’ users, focuses on the nature of that sea and the navigational procedures within it. Islands dominate the Mediterranean, and hence the sailing practices in its waters, but the Black Sea has none of any significance. So without islands that needed to be skirted round or that could serve as waypoints, Black Sea navigation is likely to have involved, for the most part, a voyage, starting at the sole entrance point, the Bosphorus, and heading more or less directly to the destination. Most of the termini would have been in the north, along the Crimean coastline or on to Tana[is] at the head of the Sea of Azov. The dangers facing pilots were also concentrated in the north, in the form of extended shoals, which the portolan charts took good care to illustrate. Nevertheless, for those heading east, say to savastopoli (Sukhumi) or trebisond (Trabzon), a dog-leg would be involved. Whether they stuck to the coast or set out to sea and changed direction at a certain point, they could not have followed a single bearing into that large, misplaced south-eastern bay. Instead, they would presumably have memorised the successive bearings involved.

Given that the reason for Mediterranean sailors venturing into the Black Sea was primarily, if not entirely, related to the various colonies and trading posts, it is likely that a correct sequence of coastal toponyms would have had more significance for the portolan-chart users than distortions in the placement of the coastlines in the south-east. That interpretation would fit in with the wider assertion that the portolan charts were strictly pragmatic navigational tools, adaptable to changing purposes and supplying precision only where it was needed.

To avoid repetition, the reader is referred for the details about the Atlantic coasts to the relevant section of the 2015 essay on the Carte Pisane. Occasional extensions into the Baltic, an area where few Mediterranean vessels ventured, provided more variety in the cartographic coverage than refinement. ³⁴⁹ 

The written coastal itineraries have already been discussed. ³⁵⁰  But what of their cartographic treatment; what might have been the source behind the outlines of the mainland and insular coasts?

It can confidently be stated that the littoral toponymic sequence is unconnected to the pelagic courses. They represent different types of information, for different purposes, and almost certainly with a different origin. However, there are competing theories about the place of the coastal outlines in the origins story, and a number of possible explanations for their source:

1. Provided, in sections, from the visual memories of mariners

Roel Nicolai has commented that:

Most scholars who endorse the medieval origin hypothesis assume the geometric basis of the charts to be a geodetic control network, consisting of the azimuths and distances between points along the coast, measured as a ship travelled along the coast (per starea) and supplemented by observations on cross-basin (per peleio) routes. ³⁵¹ 

It is possible that some kind of what could be loosely termed a coastal survey(s) might have taken place prior to the downloading of the pelagic diagram, since partial sequences of names along the coast must always have been known to those engaged in the cabotage trade. Thus adding an attempted graphic accompaniment would have been a logical step. But if that had happened the result is unlikely to have been realistic. ³⁵²  We could speculate that it might have had a similar appearance to the Charta RogerianaA 3.5 metre wall-map of the world created, in Arabic, by Muhammad al-Idrīsī in Palermo, Sicily for the Norman king Roger II in 1154, and preserved, in the form of 70 sections, in several copies of a manuscript text, i.e. with large parts unrecognisable. Hence, if the coastal outlines had come first they would have had to be heavily adjusted afterwards so as to fit into the confirmed positions of the pelagic termini.

Alternatively, the suggestion that somebody might have taken a ruler to the coastal itinerary of a portolano (Plural portolani and more usually called ‘portolans’ – which can be confused with the charts) these were medieval guides to the coastal places around the Mediterranean, listing the distances between them. They also include open-sea (pelagic) routes such as Lo compasso’Lo compasso de navegare’, c.1260-96, written in Italian and apparently designed for sailors, is much more detailed than the earlier ’Liber’ (Staatsbibliothek, Berlin, MS Hamilton 396), simply read off the distance and direction from one port to the next, and then sketched in the intervening coastal outline – as per the experiment of Jonathan Lanman in 1987 ³⁵³ – is no longer considered plausible. ³⁵⁴ 

3. Inserted later as infilling to a pelagic diagram

On the basis that the pelagic diagram came first, adding the coasts would have involved little more than filling in the gaps between those positionally-accurate termini. This is the explanation favoured in this essay. According to that contention, the chart’s fully recognisable overall outlines for the Mediterranean would be due to the relative geometric precision of those pelagic termini, and not to the drawing in of the shorelines.

As a partial corrective to that generalisation about the broad-brush approach to the mainland coastlines, Jacques Mille, who has made a close study of the Provençal coastline, found recognisable features in the 1313 atlas by Pietro VescontePietro Vesconte (fl.1311-c.30) is the first chartmaker known to us by name, and the first professional. A Genoese, he worked in Venice, along with his presumed relative Perrino. Several of his works survive, documenting in particular the growing shape and toponymy of the British Isles. His chart of 1311 and atlas of 1313 are the oldest to be signed and dated , in particular the treatment of the Giens Peninsula. ³⁵⁵  However, that is relatively close to the west coast of Italy, where the charts evidently emerged, and that area may prove to have been treated with greater care. An overall examination of the coastline treatment would be a useful subject for future research.

On our oldest surviving chart, the Carte PisaneAccepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) , the general trend of the coasts is respected, with major bays and promontories indicated, and rivers denoted with two short parallel lines. But whatever lay between was already being treated in a loose, generalised way. This left little on which to anchor the position of the ports and rivers, but neither does that seem to have been attempted.

Pietro and Perrino VescontePietro Vesconte (fl.1311-c.30) is the first chartmaker known to us by name, and the first professional. A Genoese, he worked in Venice, along with his presumed relative Perrino. Several of his works survive, documenting in particular the growing shape and toponymy of the British Isles. His chart of 1311 and atlas of 1313 are the oldest to be signed and dated began to draw capes with chisel ends, and with broader versions of that same device for the short sections between rivers. Later, for example in the work of Grazioso Benincasa An ex-ship captain, he left us a range of clearly-drawn charts and atlases (1461-82) and a written portolano (1435-45) in the mid-15th century, between the one-size-fits-all river symbols and the bays as conventional scoops, the over-emphasised headlands are conveyed by means of a variety of artificial shapes. Those are reminiscent, to us at least, of everyday objects (even if the similarities are accidental) such as tuning forks, jigsaw puzzle pieces, battlements, etc. ^355a  All this underlines the greater importance, to the charts’ creators and users alike, of the promontories (which were often pelagic termini as well), when contrasted with the casual treatment accorded to the coastal configurations between them.

I was only recently made aware, by Zoltán Biedermann, that some stretches of the coast were traditionally represented in different, specific ways. ³⁵⁶  These alternatives are already apparent in the early 14th-century atlases of Pietro Vesconte. Contrast, for example, the treatment of the uncharted north coast of Scotland, which is rendered by means of little more than a single wavy line, when compared to the intricacies of the Istrian peninsula in the north Adriatic. Other stretches fall between those two extremes. It seems likely that these distinctions had a purpose, which was probably descriptive of the differences, first, between the known and unknown, and, second, between stretches with few distinguishing features and those where rocky shorelines were present, characterised by cliffs, small inlets, reefs, and so on. ³⁵⁷ 

This aspect deserves future study to see if these distinct types of coastline representation can be classified into a typology, and, if so, how realistically that matched with reality. The value for mariners of a visual guide to the shore’s appearance when approaching an unknown coast would be obvious. It is strange that so important a feature seems to have been overlooked by all of us, past and present. However, assuming that the Carte Pisane reflects the style of its immediate antecedents, this possible further innovation must be credited to one of the chartmakers that followed later, perhaps Vesconte, and hence it has no direct bearing on the origin question itself. ³⁵⁸ 

In general, portolan charts conveyed little about a river’s size, course or source, concentrating instead on the estuaries. However, exceptions were made for the major deltas, as well as a handful of the more significant rivers, such as the Thames, Guadalquivir, Rhône, Arno, Tiber and Nile, which are slightly enlarged and extended inland. That refinement seems likely to have been designed to indicate that those rivers were navigable by seagoing vessels at least up as far as their major inland port. That, combined with the fact that few of the rivers were named on the Carte PisaneAccepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) , might seem to indicate that the portolan charts’ initial focus was on seaborne rather than riverine trade, often naming the ports at their entrance rather than the river itself.

Dating probably from a little after the Carte Pisane, the Cortona chartBased on its archaic appearance, this is considered to be the second oldest surviving portolan chart after the Carte Pisane, dating from the beginning of the 14th century or even earlier. It was first described in 1957 (Biblioteca dell'Accademia Etrusca di Cortona, Port. 105) introduces inland detail for the Danube and Nile, as well as a river at Latakia (liche), Syria. ^358a  By 1313, Pietro VescontePietro Vesconte (fl.1311-c.30) is the first chartmaker known to us by name, and the first professional. A Genoese, he worked in Venice, along with his presumed relative Perrino. Several of his works survive, documenting in particular the growing shape and toponymy of the British Isles. His chart of 1311 and atlas of 1313 are the oldest to be signed and dated was using island-like segments, in different colours, to symbolise the complexity of such major deltas as the Danube, Dnieper [Dnipro] and Nile. The Danube is the second largest delta in Europe but is indicated on the charts by no more than the two main outlets into the Black Sea at that time, the Sulina and Stantul Gheorghe. This reticence was presumably a response to the shifting channels and large areas of marshland involved. As with other symbols, their meaning must have been left to oral explanation.

However, many rivers were navigable, if only via transhipment into smaller vessels. In those cases the inland ports might be named even if some, for example Cologne, Paris or Seville, were a long way upriver. That information would clearly have been important for merchants but it was only the Catalan-style charts from 1330 onwards that introduced generalised courses for the major rivers, as well as the cities along them, thus indirectly offering information about upstream navigation. ³⁵⁹ 

It might have been expected that the quick and easy method of drawing the coast as a single continuous line would have been adopted. Instead, as can already be seen on the Carte Pisane, pairs of short parallel lines were used to denote river estuaries. This was a formalised symbol that took no note of the difference in size between the medium and small rivers. That ingenious, minimalist device has not apparently been noted in earlier cartography. Because the ink line turns briefly inland and then stops abruptly, to be taken up again for the start of the next stretch, the draftsman had to raise his pen and then replace it, so as to indicate the river’s other shore, before continuing along the coast. Like the intricate symbols for offshore dangers ³⁶⁰ no effort seems to have been made to find drafting shortcuts. ³⁶¹  However, there might be one exception, although this point was not thoroughly checked, because some of what would become river mouths on the Vesconte charts had previously appeared as small inlets on the Carte Pisane and Cortona chart (i.e. they had initially been drawn as part of the continuous coastal outline).

K.1d. Islands

For anyone navigating the northern half of the Mediterranean, islands played an important role, since they provided copious reference points to help fix a ship’s position. ³⁶²  The way that the charts handled the islands of different sizes gives us clues as to how that information might have been used by mariners and, at the same time, how it was dealt with by the chartmakers.

As to their outlines, five of those larger islands had already reached what was effectively their final forms on the early work of Pietro VescontePietro Vesconte (fl.1311-c.30) is the first chartmaker known to us by name, and the first professional. A Genoese, he worked in Venice, along with his presumed relative Perrino. Several of his works survive, documenting in particular the growing shape and toponymy of the British Isles. His chart of 1311 and atlas of 1313 are the oldest to be signed and dated (1311-13). Only Majorca and Crete had to wait until the 1330s for the improvements seen on the charts of Angelino DulcetiAngelino Dalorto or Dulceti was a Genoese who worked in Palma, Majorca, leaving us dated charts of 1330 and 1339. His were the first to introduce inland and illustrative details. Those outlines are always recognisable and, in the case of Sardinia, Sicily and Crete, notably realistic, so much so that there would only be modest improvements later. ³⁶³ 

The first hints of what would prove to be an unusual refinement in the portrayal of the smaller islands can be seen already in the Carte Pisane’sAccepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) treatment of one or two of the Aegean islands, for instance, Skyros (squur?) and Amorgos (mergo), which is shown as a castellated wall. These are a foretaste of the distinctive, and wholly imaginary, outlines, which would, in the work of Vesconte and Angelino Dulceti be developed into fully-fledged ‘signature’ shapes [Powerpoint, Slides 18 & 19]. Those were not facsimiles of reality but rather specific, easily memorised logos, which were found to be useful enough to be generally adopted and then repeated for centuries. Some island groups elsewhere were treated similarly, particularly those along the eastern Adriatic. This conceptual leap, which would be fully developed as a mnemonic device only some decades after the Carte Pisane, is otherwise unknown in the history of cartography. ³⁶⁴  It is unlikely that the architects of the Dubai ‘Universe’ were aware that they were not the first to invent island shapes [Powerpoint, Slide 20].

From the outset, the portolan chartmakers took great care to label each island with its toponym. Sometimes, because the coastlines needed to be left free for the depiction of marine hazards, the toponym of a nearby island or offshore danger might be inserted into the sequence of coastal names. Sailors would have had to learn to identify which offshore feature was being referred to in each case. ³⁶⁵  This is a further indication of the importance seamen assigned to islands and, likewise, to marine hazards.

I owe to Juan Ceva the suggestion that there could be another explanation for distorting island outlines. Studying the Balearic Islands, he had observed that the natural harbours had been exaggerated in size, presumably to draw attention to safe anchoring places. {This paragraph added 9 November 2023.}

K.1e. Charting the Aegean archipelago

Pelagic courses in the Aegean – for example those described in the LiberThe ’Liber de existencia riveriarum’, c.1210, written in Latin, is the oldest surviving guide to the Mediterranean, giving the distance and direction between places along the coast and across the open sea (Cotton MS Domitian A XIII, ff. 114r–129v) and depicted in the diagram at the end of Patrick Gautier Dalché’s edition (1995) [Powerpoint, Slide 11] as well as Joaquim Alves Gaspar’s 2019 article ³⁶⁶ – were, of necessity, relatively short and usually broken into sections. Those would not have provided a reliable underpinning for a chart of the Aegean in the way this essay is claiming that they did for the Mediterranean.

Since the Aegean could not have been visualised as a whole, it seems inescapable, in this case at least, that something akin to an actual survey must have been carried out. To have plotted the entire network, as seen on the Carte PisaneAccepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) , it would have been necessary to locate the position of every island relative to the others (even if only approximately) and, where relevant, to the mainland coastline as well. Whereas the mental-map origin theory proposes that the Mediterranean coastlines and toponymy were later infillings to the web of pelagic courses on the true (lost) urdocument, ³⁶⁷ in the case of the Aegean that proceeding must have been reversed. The detailed data would need to have been collected first, at the level of small groups of islands which could then be joined up.

It is not suggested that anybody would have systematically set out to capture the intricate patterns formed by the overall archipelago, and they would certainly not have been motivated by anything approaching a modern concern for accuracy and completeness. Nor is it likely that a single individual would have visited all the islands; any more than the author of the earliest isolario, Cristoforo Buondelmonti, was engaged in a primarily geographic survey when he gathered material for his Liber insularum arcipelagi around 1420. ³⁶⁸  The more likely scenario would have involved the collaborative sharing and combining of overlapping personal experiences, as well as the adjustment stage that must have been necessary to reconcile contradictions and correct errors (though not all of them).

The distortions in the overall shape of the Aegean’s coastline, and some imprecision in the placement of the individual islands that can be seen on the Carte Pisane, are the inevitable result of the inherent difficulties, as well as the limited means at a medieval pilot’s disposal. Despite those restraints, the importance of the Aegean was such that some portolan charts, among them the oldest surviving atlas, Pietro Vesconte’sPietro Vesconte (fl.1311-c.30) is the first chartmaker known to us by name, and the first professional. A Genoese, he worked in Venice, along with his presumed relative Perrino. Several of his works survive, documenting in particular the growing shape and toponymy of the British Isles. His chart of 1311 and atlas of 1313 are the oldest to be signed and dated of 1313, treated it separately, and at a larger scale than the rest. ³⁶⁹  Overall, the portolan charts’ survey of the Aegean is remarkable, and one that can be contrasted with the Charta RogerianaA 3.5 metre wall-map of the world created, in Arabic, by Muhammad al-Idrīsī in Palermo, Sicily for the Norman king Roger II in 1154, and preserved, in the form of 70 sections, in several copies of a manuscript text, which shows around 30 of the 100 islands and islets (most with their names), but presented with non-specific shapes and with only a minor range in their sizes. ³⁷⁰ 

The specific characteristics of the Aegean invite speculation as to how medieval pilots navigated through it. Besides those ships sailing specifically to the numerous (mostly Venetian) colonies, strongholds and trading posts, there was a steady flow of vessels making their way to and from the Black Sea. No doubt some repeated voyages would have followed standard routes, at least when the expected wind was blowing. But how might the details of a preferred, necessarily zigzag, course have been mentally retained by the helmsman? It is fair to assume that he would have stored in his memory pictures of the distinguishing appearance of the flank of each island he would be passing. He would remember each by its name and, from about 1320, by its mnemonic ‘signature’ shape and colour, ³⁷¹ and recall which had to be left to port and which to starboard, as well as the identity of the ones that loomed up ahead. Because there is no direct route through the archipelago, the frequent changes of direction (recorded in terms of the wind compass) would also have formed part of the navigator’s composite memory and thus helped to confirm the relative positions of those islands.

Although the portolani, and particularly Lo compasso’Lo compasso de navegare’, c.1260-96, written in Italian and apparently designed for sailors, is much more detailed than the earlier ’Liber’ (Staatsbibliothek, Berlin, MS Hamilton 396), might refer to those dangers and even describe their general nature and approximate location, only a graphic document could indicate both aspects with any precision. The Carte PisaneAccepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) , along with the unsigned charts that followed it and the earliest productions of the VescontePietro Vesconte (fl.1311-c.30) is the first chartmaker known to us by name, and the first professional. A Genoese, he worked in Venice, along with his presumed relative Perrino. Several of his works survive, documenting in particular the growing shape and toponymy of the British Isles. His chart of 1311 and atlas of 1313 are the oldest to be signed and dated workshop, include the first graphical representations of several of these features. ³⁷⁵  These would be steadily developed during the 14th century, incidentally providing further evidence of the chartmakers’ ingenuity.

The portolan chart provided the only effective means of conveying a category of information that was vital for sailor’s safe-keeping. Translating an oral description of what had been directly experienced into visual form was a far greater challenge for chartmakers than introducing extra place-names. Nevertheless – and this remains to be tested in detail – the steady introduction of new hazards onto the charts seems to have matched the introduction of new toponyms. The portolan charts were never more relevant than when they added details of localised hazards, allowing those to be widely shared.

K.2a. Earliest mentions

In the context of the charts’ origin it can be noted that the Skerki Bank, to the south-west of Sicily and probably the single most treacherous feature in the Mediterranean, is named and depicted on the Carte Pisane, as well as being referred to in both the LiberThe ’Liber de existencia riveriarum’, c.1210, written in Latin, is the oldest surviving guide to the Mediterranean, giving the distance and direction between places along the coast and across the open sea (Cotton MS Domitian A XIII, ff. 114r–129v) and Lo compasso’Lo compasso de navegare’, c.1260-96, written in Italian and apparently designed for sailors, is much more detailed than the earlier ’Liber’ (Staatsbibliothek, Berlin, MS Hamilton 396) [Powerpoint, Slide 24]. The Carte PisaneAccepted as the oldest surviving portolan chart, c .1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) also shows, prominently, two offshore features to the south and east of the Skerki Bank: the Bancs de Kerkennah (Tunisia) and the Shoals of Sirte (Libya), the latter much feared by the ancients.

It seems likely that some such warnings to sailors would also have been present, albeit in rudimentary form, on the antecedents of the portolan chart, since they would have been preserved in sailors’ memories in the same way as the pelagic courses. The earliest known unambiguous reference to the use of the portolan charts at sea, also emphasises the importance of the warning they carry about marine perils. In Egidio Colonna’s De Regimine Principum (1279), he explained that sailors:

seeing the dangers of the sea and wanting to avoid wrecking their own ships, draw a map of the sea on which sea ports, hazards and other such things are described in proportion, from which sailors, looking at them attentively, immediately perceive where they should go and where they are and in which places they should be careful. ³⁷⁶ 

K.2b. The treatment of dangers in the portolani and chart compared

The incorporation into a pilot guide of references to offshore dangers was already more complicated than the gathering of toponyms, or the simple statements of distance and direction from one cape, harbour or island to another. Whereas the early written guide could tell the sailor roughly how far it was, say, from Tunis to southern Sardinia, and indeed the initial course to steer, it would not help him avoid the wreck-strewn barrier of the Skerki Bank [Powerpoint, Slide 24].

Patrick Gautier Dalché identifies in the LiberThe ’Liber de existencia riveriarum’, c.1210, written in Latin, is the oldest surviving guide to the Mediterranean, giving the distance and direction between places along the coast and across the open sea (Cotton MS Domitian A XIII, ff. 114r–129v) a reference to that feature, which is there given the early form of its name, chilbi. ³⁷⁷  Though the nature of the danger is left unspecified, it is placed, with reasonable accuracy – in terms of their respective directions and distances in 'miles' [milliara] – in relation to Trapani on Sicily's west coast and Isola Marettimo to the west of that, as well as to Tunisia's Cape Bon and the island of Gimari [Djamour El Kébir] in the Gulf of Tunis. ³⁷⁸  Likewise with Lo compasso’Lo compasso de navegare’, c.1260-96, written in Italian and apparently designed for sailors, is much more detailed than the earlier ’Liber’ (Staatsbibliothek, Berlin, MS Hamilton 396), it is theoretically possible to triangulate the various supplied directions between what it terms the quilbo shoal, to ports, headlands and islands in its vicinity. But would anybody have wanted to comb through the ten dispersed references to the Skerki Bank in that un-indexed text to determine those facts? ³⁷⁹ 

Indeed, it can plausibly be argued that among the likely reasons for the charts' creation in the first place was the plotting of those navigational dangers, in a way that revealed their location, extent and nature. In the time-and-money equation for a master chartmaker, the sheer number of those warning symbols – perhaps as many as 3,500 ink crosses or separate dots on a single chart or atlas ³⁸⁰ – and the care with which they were precisely replicated over centuries, must have represented a significant outlay, thus underlining the importance accorded to that feature.

Kevin Sheehan, an experienced cartographic draftsman, provides useful information in his 2014 doctoral thesis about the way that the charts were drawn. ³⁸¹  As a result of his carefully documented copying of three portolan charts, according to what seemed to be the most likely contemporary methods, he was able to estimate that producing an undecorated chart of the full usual extent would have taken around three weeks. ³⁸²  Breaking down the separate elements of his chosen model – which excluded the Atlantic, eastern Mediterranean and Black Sea – he found that the rhumbline network took about four hours to copy, the ‘hydrography’ and toponymy about twelve each and the combined scales, colour and ‘other’, 11.5 hours. ³⁸³  Unfortunately, it is not possible to extract the danger symbols from those figures because those in black ink were subsumed into the hydrography total and the red dots are included with the red-inked names in the toponymy figure. Nevertheless, it is likely that the time spent on depicting the offshore hazards would have been comparable to that devoted to the outlines and toponymy.

The Carte PisaneAccepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) includes a limited number of those warnings, whereas Lo compasso is liberally provided with them. Several of those referred to on the latter are not seen on charts until considerably later. In a detailed study of France’s south coast on the Avignon chartA chart fragment focusing on the North Sea, probably very early 14th century, discovered in 2002 in the Vaucluse Archives (Port 01, 3E 54) – for Jacques Mille’s studies on this see the Portolan Chart Bibliography: https://www.maphistory.info/portolanref.html , Jacques Mille found that some of Lo compasso’s descriptions were as good as their first cartographic depictions on the 15th-century charts of Francesco and Batista Beccari Francesco and his son Battista were active in or around Genoa in the first half of the 15th century. Francesco’s 1403 chart is notably innovative (New Haven, Yale University, Beinecke Library). ³⁸⁴  It is not clear though how that portolano’s author was able, in those cases at least, to pre-empt the chartmakers by well over a century. ³⁸⁵ 

K.2c. How were the dangers recorded and how did those get onto the charts?

The study carried out by Michael Barritt and myself identified 96 clear threats to navigation (both then and now), almost all of them named. ³⁸⁶  54 of those were noted on the portolan charts. The remainder, instead or as well, can be seen in the 13th-century portolani (Singular portolano and more usually called ‘portolans’ – which can be confused with the charts) these were medieval guides to the coastal places around the Mediterranean, listing the distances between them. They also include open-sea (pelagic) routes or their successors. Carte PisaneAccepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) already includes 22 of the 54 hazards found on the charts. Additions to that number were made in the 14th century and a further 20 after 1400. ³⁸⁷  Even though some of the later introductions can be shown to be imaginary, taken together they represent perhaps the most significant updating to the charts, if toponymy is left to one side.

Although those dangers would certainly have had local names bestowed on them, most are depicted anonymously on the charts. In Lo compasso’Lo compasso de navegare’, c.1260-96, written in Italian and apparently designed for sailors, is much more detailed than the earlier ’Liber’ (Staatsbibliothek, Berlin, MS Hamilton 396) they are usually referred to in generic terms, as scoglie (rocks or reefs) or seche (shoals). ³⁸⁸  This lack of an identifying name may help to explain why these hazards have been little discussed by historians. Yet 43 instances are referred to by name in Lo compasso – as distinct from just thirteen in the earlier LiberThe ’Liber de existencia riveriarum’, c.1210, written in Latin, is the oldest surviving guide to the Mediterranean, giving the distance and direction between places along the coast and across the open sea (Cotton MS Domitian A XIII, ff. 114r–129v), whose clerically-influenced author evidently considered navigational information to be of less interest. ³⁸⁹  Attempting to match those brief textual mentions with what can be recognised on today’s marine charts could be a valuable project for future research.

Confusingly, a number of the dangers that were indeed named on the charts were placed by the draftsmen in the general toponymic sequence, and thus, at first glance, presented as if they were land features. ³⁹⁰  Absence of space in the sea nearby might sometimes have been the reason for that, though not generally. But are those 'inappropriate' names 'interlopers' or do they rather point to the need to see the charts through their users’ eyes and thus upgrade the contemporary importance attached by them to such offshore features, sufficiently well known for their names to have been squeezed into the coastal sequence?

Closely related to the above are the lighthouses and other devices designed both to signal danger and help mariners avoid them. The word ‘pharos’ was the medieval term for a lighthouse and about ten instances of faro have been noted in the portolan chart toponymy. ³⁹¹ 

Related to the issue of what features could jeopardise a ship's safe passage, and reiterating how that would be a central concern for any medieval sailor, it is reasonable to expect that attempts would have been made, probably routinely, to seek information orally from local experts. I owe to Michael Barritt a suggestion, which has not apparently been clearly articulated before, about the potential role of fishermen as informants. In his own career and certainly in those of his predecessors as hydrographic surveyors, they would, as a matter of course, have consulted those with the best local knowledge. He cites the experience of Captain William Durban of Weazle, who, during his survey of the Skerki Bank in 1802, called in at Trapani, Sicily, where he embarked some local fishermen. On his return to the bank he almost immediately found its shoalest point, to which he gave its present name, Keith Reef. ³⁹² 

If that suggestion is accepted, fisherman – various and anonymous – would have a real claim to be one, if not the main, source for information about the in-shore navigational dangers that form so important an element of the portolan charts' purpose and usefulness. If so, that contribution has hitherto been ignored. That Francesco BeccariFrancesco and his son Battista were active in or around Genoa in the first half of the 15th century. Francesco’s 1403 chart is notably innovative (New Haven, Yale University, Beinecke Library) does not mention them on his chart of 1403 in the roll-call of those who contributed updated information for the charts – “masters, shipowners, skippers and pilots of the seas of Spain and those parts” – is unsurprising. ³⁹³  He was not a sailor and, as a chartmaker, would have received any information from such a source via an intermediary, possibly without realising how it had been obtained.

Those on board a medieval vessel, whether mariners, merchants, soldiers, travellers or pilgrims, went in dread of the very real natural hazards they faced. These ranged from the vulnerability of their ships to weather events, hazards lurking offshore, or the threat from pirates, and this led naturally to prayer. ³⁹⁴  As Aaron J. Brody pointed out “Maritime religions were a discrete subset of the general religious beliefs and cultic practices of ancient Mediterranean societies, generated by the unique uncertainties and dangers faced while living and voyaging at sea”. ³⁹⁵ 

In this context, promontories played a special part, with the charts aiding the mariners’ search for divine assistance. A number would have had a temple shrine visible from the sea but in addition, as Christer Westerdahl explained, promontories “were also dangerous, not only as cliffs to founder on in unfavourable winds, but often exceptionally so, since capes often mark a change in climatic conditions with conflicting winds, and expose seafarers to strong and sometimes fickle currents, sometimes at the same time”. ³⁹⁶ 

Leaving geographical boundaries aside, what is likely to have been of greater relevance to sailors would have been those of saintly suzerainty, and the charts’ place-names would have enabled them to identify the appropriate saint for that locale, who could be asked to intercede in times of special danger. ³⁹⁷  After all, it was God who had the power to save or abandon the mariner when at sea, not the local lord.

Among those who would have benefited from the 2,000 or so place-names on a standard chart were non-sailors with a strong connection to the sea, particularly shipowners and merchants. But the portolan chart did not come into existence to service their needs and no features have been observed that would have been added specially for them ⁴⁰¹  Nevertheless, the charts’ impressive body of littoral and island place-names, and the visible relationship of the various ports, would have been valued by the mercantile class, who would now have a better understanding of the overall dispositions, particularly when those related to the voyages of their own vessels. ⁴⁰² 

In seeking to understand the relationships between, respectively, the Carte PisaneAccepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) , the LiberThe ’Liber de existencia riveriarum’, c.1210, written in Latin, is the oldest surviving guide to the Mediterranean, giving the distance and direction between places along the coast and across the open sea (Cotton MS Domitian A XIII, ff. 114r–129v) and Lo compasso’Lo compasso de navegare’, c.1260-96, written in Italian and apparently designed for sailors, is much more detailed than the earlier ’Liber’ (Staatsbibliothek, Berlin, MS Hamilton 396) (and particularly the first of those), various questions need to be addressed. How might that toponymic information have been gathered? Could either the chart or the written guide have been the primary source for the other? In other words, what might have been extracted from a chart for reuse in a text; or, alternatively, how could a chart’s toponymy have been populated, and the individual names positioned, using a text? The succession of places and features along the coast would surely have been absorbed by a mariner through experience, rather than being learnt from a portolano (Plural portolani and more usually called ‘portolans’ – which can be confused with the charts) these were medieval guides to the coastal places around the Mediterranean, listing the distances between them. They also include open-sea (pelagic) routes , and with the appearance and characteristics of each carefully filed away in memory. Passing from one harbour to the next could take a matter of hours, and, with nothing else to look at, there would have been plenty of time to record the visual details.

Against that, given that its toponymy is the only textual element on a functional chart, it is not unreasonable to assume that the place-names would have derived from a written source. In that scenario, a pilot guide, with its littoral itinerary, would be the obvious candidate. To test that, a comparative analysis was made of the place-names on the two portolani and the Carte Pisane. In the first place, little of value would have been learnt by considering, for example, places such as Barcelona, Dubrovnik or Sfax, since those would inevitably have been included. A more relevant comparison is one that focuses on the less common, even unique names. So, considering only toponymic incidence, rather than the erratic name forms – in other words treating the presence of a toponym merely as a measure of what was both known to, and sufficiently valued by, the respective compilers – the standard names were then removed from the analysis. 'Standard' is defined as those that appeared among the roughly 15% that were highlighted in red ink throughout the life of the portolan charts.

There were two further refinements. First, the analysis was restricted to the Mediterranean because the Atlantic and Black Sea coasts were not covered in all three of the documents being studied and, second, because the underlying toponymic analysis ⁴⁰³ had concerned itself with the mainland names only, the island names were also omitted. The resulting overlap between the toponymic lists in the Carte Pisane and the two portolani would then reveal the convergence or individuality of each, in a more meaningful way. ⁴⁰⁴ 

The result showed significant differences in the three toponymic selections ⁴⁰⁵ [Powerpoint, Slide 25]. This was surprising, given that it is reasonable to have anticipated that anyone who sailed widely in the Mediterranean would have taken note of many, if not most, of the same features, whether those were ports, river estuaries, prominent capes and bays, landmarks that could be used to fix the vessel's position (such as churches, castles or mountains), or offshore islands.

From the percentages of the shared non-standard names it is clear that neither the Carte Pisane nor the other very early surviving charts were in any meaningful sense derived from the antecedent portolani texts. Probably decades of experimentation preceded the Carte Pisane but the two early portolani are unlikely to help us visualise the toponymy of such hypothetical precursors. Furthermore, turning to the two texts, no more than 60% of the Liber’s non-standard place-names can be found in Lo compasso – despite an attempt to match up the Liber’s Latin forms with the Italian versions in Lo compasso. ⁴⁰⁶  This reinforces the significant differences between the Liber’s names, which are listed in an anti-clockwise direction, compared to the clockwise preference of Lo compasso’s author. In its turn, the Carte Pisane’s debt to each of those texts is less than 45%. The dissimilarities of the littoral itineraries reflected in the toponymic sequences found on the Liber and Lo compasso on the one hand, and on the Carte Pisane on the other, suggest that at least three, unconnected, comprehensive toponymic surveys were made during or before the 13th century, unless future research found that all three might be interpreted as selections from a far more detailed lost original. ⁴⁰⁷  However, that seems unlikely because, whereas a portolan chart is restricted in the number of names that can be squeezed in, the same would not have applied to a text.

Further evidence that the portolani did not serve as inspiration for the charts comes from the context within which the toponymy was presented and the different purposes those imply. For the portolani, statements of the route from one harbour or headland to the next gave equal weight to distance and direction. Yet the charts, while respecting the coastal sequence (apart for occasional confusion along the Dalmatian shoreline), showed limited interest in the respective distances between the toponyms. Hence the portolani could not have derived their coastal itinerary from the charts because their imprecise placement of the toponyms meant that distances could not have been calculated from those.

Not only were there different origins for the place-names in the written narrative and the drawn chart but, in certain respects, their toponymy remained divergent over the centuries, with a number of names being unique to one or other format. ⁴⁰⁸ 

Leaving aside the coastlines of the north-east Atlantic and North Sea (whose addition to the charts only occurred after the Carte PisaneAccepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) , ⁴⁰⁹ as well as the Black Sea, the two earliest surviving portolani (Singular portolano and more usually called ‘portolans’ – which can be confused with the charts) these were medieval guides to the coastal places around the Mediterranean, listing the distances between them. They also include open-sea (pelagic) routes and the Carte Pisane are more or less complete in their geographical coverage, without leaving any long, empty stretches.

The portolan charts’ mix of (often unstated) referents for their toponyms is probably unprecedented. This embraced a mélange of coastal and offshore features (even occasionally those well inland) such as gulfs, ports, capes, nearby islands, churches, castles, offshore dangers, rivers, lagoons, salt flats and mountains. ⁴¹⁰  This ambiguity, helped by the lack of the standard cartographic black dot alongside each toponym, made it easier for the charts’ user, whether intentionally or not, to attach the name to whatever referent – perhaps more than one – he wished. Take agulones, for example, introduced just north of La Rochelle by the Pizzigani brothersThe Venetian chartmakers, Domenico and Francesco, left us a signed chart and an atlas from, respectively, 1367 and 1373 in 1373. It may indicate the port of L'Aiguillon, the Pointe de l'Aiguillon (a cape), or Anse de l’Aiguillon (an inlet). Or perhaps all three. It is equally possible that the meaning of a toponym itself might change over time. ⁴¹¹ 

Unless, that is, the feature being referred to was specifically identified by means of a prefixed initial: c (capo, point, promontory or headland), g (gulf or bay), i (isla or island) or p (porto or punta – ambiguously, port or headland) – with rarer occurrences of mons (mountain when of significance for navigation), or words pointing to offshore reefs or shoals [see the previous section on ‘Navigational Dangers’]. Rivers might be denoted by a preceding r or a suffix fl. (fluvius). But frequently there was no epithet. Occasionally, the sequence includes the name of the region or even the country. The Carte Pisane Accepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) ’s toponomy for the North Sea coast, for example, lists in turn, and in a consistent lettering style: allamaigna, flandis, brugis, porto niua, gravalingue, friza (Germany, Flanders, Bruges, Nieuwpoort, Gravelines, Friesland). For those cases, and the instances where no clarifying term was provided, comprehension depended on the reader’s existing knowledge, backed up perhaps by orally conveyed interpretation from others.

For portolan chart toponymy to be properly understood by modern historians it often needs to be interpreted as indicative rather than precise, both as to its potentially dual or multiple meanings and its geographical placement, i.e., more of a label to jog memory than a unique identifier. This ambiguity also had its advantages for the copyist, since, by leaving the toponym with its multiple potential meanings, this simplified his work, and meant that – even where there was sufficient room – repetition could be avoided.

Maybe we should interpret those single names, which could, potentially, have had multiple references, as portmanteau terms for any or all of the local applications to which the sailor's knowledge or memory could be attached. And might such an ambiguous toponym have sometimes also highlighted an offshore danger, which was known to them if not to us, by the name of a nearby feature or settlement?

K.3d. Red names

The earliest surviving chart (the Carte Pisane Accepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) ) already reveals its debt to monastic illumination in its use of red ink to emphasise the names of the more significant ports; though we should not read into that any ecclesiastical involvement in the charts’ origin. Those red names would later constitute approximately one-sixth of the total, as more toponyms were ‘promoted’ to red. Surprisingly, a large number of chartmakers were apparently the first to do that in at least one instance. ⁴¹²  That fits in with the parallel situation in which some of the place-names written in black also reflect the choice of a separate chartmaking school or even an individual.

About 90% of the red names on a portolan chart refer to ports, whereas the geometric structure of the pelagic courses had focused on the capes. This underlines the separate development of the charts’ graphic and textual elements. Natural features, however prominent, were almost never treated in red on the Carte Pisane. ⁴¹³  This might seem to conflict with the sailors’ focus on headlands. Two possible reasons can be offered: that treating a cape and major port in the same way would have been confusing, and, further, that too much red – given that most promontory names brought with them a prefix, cauo de – would have diminished its effect. ⁴¹⁴ 

K.3e. Updating the names

Since we have no means of knowing how the charts’ toponymy might have developed between their inception (probably in the late 12th century) and the Carte Pisane Accepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) , and since the names in the two 13th-century portolani (Singular portolano and more usually called ‘portolans’ – which can be confused with the charts) these were medieval guides to the coastal places around the Mediterranean, listing the distances between them. They also include open- sea (pelagic) routes are markedly different ⁴¹⁵ we must look to the post-1270 developments on the charts themselves, for clues as to what might have gone before.

The place-names on a portolan chart were usually written with care. As a result, they are, for the most part, fully legible even on a chart where the smaller-than-usual scale reduces the lettering to minuscule proportions. Since they broadly adhere to a standard sequence, they can be readily documented and corralled into comparative lists. In addition, the general recognisability of the toponyms contrasts with the fluidity and subtlety of the coastal configurations. Moreover, despite what had earlier been assumed otherwise, portolan toponymy has proved to be highly dynamic. ⁴¹⁶  For those reasons, their place-names have been the most closely studied feature on the charts.

The appearance in 2007 of the DVD accompanying Ramon Pujades’s Les cartes portolanes: la representació medieval d'una mar solcada, ⁴¹⁷ offering enlargeable and rotatable scans of almost all portolan charts and atlases up to 1469, made possible the extensive comparative study of the toponymic component of charts produced by different individuals, schools and periods. The present author’s comprehensive listing of the mainland names from Dunkerque to Essaouira (Mogador ) indicates (among several other details) when each name was first noted on a dated or anonymous chart. ⁴¹⁸  The impressive number of new entries – many hundreds added over the centuries to the initial 1,000 seen on the earliest charts of Pietro VescontePietro Vesconte (fl.1311-c.30) is the first chartmaker known to us by name, and the first professional. A Genoese, he worked in Venice, along with his presumed relative Perrino. Several of his works survive, documenting in particular the growing shape and toponymy of the British Isles. His chart of 1311 and atlas of 1313 are the oldest to be signed and dated – might lead to an assumption that there was a clear and continuing programme to update that element on a regular basis. ⁴¹⁹  The fact that each of the identified chartmakers working up to 1408 apparently contributed at least one new mainland name to the total would appear to support that assumption. ⁴²⁰ 

However, that observation provides a partially misleading impression of currency. A better guide to the development of portolan-chart toponymy can be found in the detailed analyses that have been published in recent decades. For example, in the case of the few settlements whose naming or foundation is firmly attested, it can be shown that it took an average of 75 years before their first appearance on a surviving portolan chart. ⁴²¹ 

More generally, it is abundantly clear that portolan chart toponymy was not systematically kept up to date. That nearly 800 mainland names were added to the initial 1,000 between the time of the first charts by Vesconte (1311 and 1313) and 1600 ⁴²² does not contradict the fact that very few of them were new toponyms, although some might have become newly relevant in the eyes of mariners. This might indicate that sailors had memorised much of that information anyway and were thus already aware of most of the omissions that were going to be rectified later. But that also meant that medieval mariners could not have relied on the topicality of the place-names in the way that modern assumptions might expect. If toponymic currency had been important to them, then surely they would have demanded regular updating before paying the considerable sum for a chart they would have considered to be at least partially obsolete. Even the sizeable group of Venetian sailor-chartmakers in the 15th century contributed little to the charts’ toponymy. ⁴²³ 

Another consideration that should not be ignored relates to the mechanisms for conveying the details of new or altered toponyms to an individual chartmaker and, through subsequent imitation, to others. Clearly, some of the additional names could have been supplied directly to a chartmaker by a passing sailor from the locality concerned. But the innovations in question are spread across the entire Mediterranean and Black Sea so that explanation is inadequate. Instead, for a fresh name to be added to a chart three distinct stages might have been involved:

• The name was supplied by local informants: perhaps harbour-masters and seamen met in the port or fishermen operating from there
• The name was remembered and carried back by an intermediary – perhaps a seafarer who gathered omitted names when he travelled abroad – in the form that locals told him was then in use
• The information would finally be transmitted to the selected chartmaker, who, in the early years, would have been based in an Italian port or Majorca. It was left to him whether to include it on his charts or not. Even if he decided to add it, the name’s perpetuation would depend on other chartmakers imitating him ^423a 

If that tentative reconstruction is correct it would mean that three oral/aural stages were involved, unless an educated member of the ship’s company had written the information down instead. This would have been a far from straightforward process. What was said by the local informant might be misheard. Even if remembered correctly, the same could happen when the toponym was passed, perhaps months later, to the chartmaker. He, in turn, might misunderstand or intentionally adapt what he had heard to suit his linguistic preferences. Finally the names that had been accepted were sold back to mariners in the form of a working portolan chart. Even if that suggested sequence refers to the updating of the toponymy on later charts, a similar procedure might well have played its part from the outset. It is worth remembering that the portolan-chart toponymy would have been fully up to date only at the time the very first charts were being constructed; thereafter, it was left to accidental, uncoordinated updating {this sentence added 26 February 2021}.

Francesco BeccariFrancesco and his son Battista were active in or around Genoa in the first half of the 15th century. Francesco’s 1403 chart is notably innovative (New Haven, Yale University, Beinecke Library) is a notable exception to the general laissez-faire attitude of chartmakers towards toponymy. His sole surviving chart of 1403 is notable in many ways, one of them being its large number of new place-names. His statement on that chart that he had relied on a range of informants – ‘masters, shipowners, skippers and pilots’ – indicated that he actively sought to update his own toponymy. ⁴²⁴  That his additions were slow to be imitated by rival chartmakers is not untypical. Clearly chartmakers would have seen each other’s productions – shown to them by sailors no doubt – but they were apparently selective about which additional toponyms, if any, they would copy onto their own charts. This downgrading of the importance of place-names should be seen as both cause and effect of the erratic and usually delayed incorporation of new information onto the charts. ⁴²⁵ 

Portolan-chart toponymy has served as the focus for many studies by cartographic historians seeking evidence for attributing unsigned works to a chartmaking school or individual. Further detailed comparative studies of the names on the Carte Pisane and its first successors might well help to clarify the different lineages there could have been in the period leading up to that first surviving chart.

The merchants and shipowners, who are likely to have relied on the charts for the names and location of ports and islands, seem not to have acted as a stimulus for toponymic updating. Perhaps, they were simply unaware of the charts’ limitations, or considered those to be insignificant.

Looking at the wider context, it can be noted that neither the original selection of names nor those added later were erudite or included for their historical significance. Clearly, no savants were consulted. Nor, with the exception of the Holy Land with the dense concentration of red names along its coast, was the toponymic sequence designed specifically for travellers. The occasional ‘vecia’ (e.g. Alexandria) is surely more likely to have been noted because the ruins were visible from the sea. Essentially the dynamism of the charts’ toponymy reflected its deeply democratic, crowd-sourced origin managed entirely, as Beccari explained in 1403, by and for the seafaring community. {This paragraph added 9 November 2023.}

It seems likely that the secondary importance attached to updating the toponymy through the long life of the portolan charts could well have applied at the time of their origin. As already suggested, the geometric diagram that was likely to have preceded the first recognisable chart would probably have restricted its initial toponymy to those features and places that had been selected as pelagic termini, meaning that what were considered to be less significant intervening toponyms would only have been added later as the portolan chart evolved. ⁴²⁶ 

The increased cost of those added features that had no relevance for navigation point away from shipboard use towards merchants’ offices or the homes of the wealthy, even rulers. Such adorned charts survive in vastly disproportionate numbers compared to the very rare working versions. ⁴³⁰  Virtually mass produced, those functional examples, which would have been standard equipment on board a ship, have almost entirely disappeared (except for a few scattered fragments retrieved from archival bindings) having presumably been worn out through use and then discarded. ⁴³¹ 

Maritime lore at the conjectured time of the first portolan chart (c.1270) was spoken, heard and assimilated, demonstrated and remembered, but not written down. As Joaquim Alves Gaspar summarised: “There is no doubt, whatsoever, that navigation prior to the chart era – both coastal and oceanic – was conducted on the basis of the knowledge and experience of pilots: about courses, distances, navigational dangers, prevailing winds, etc.” ⁴³² 

As a result, mariners must have carried in their heads three things: a memorised databank (made up of course instructions and marine lore generally); what might be termed an ‘imagebank’ (for recognition snapshots); and the equivalent of a video archive (for instance, when moving through the Aegean). Nor can the portolan chart itself be understood without accepting the hugely important role of orality in its origin and use. The wholly pragmatic, artisanal environment out of which we are claiming the chart emerged can only be appreciated with an orality-gaze, not a literary one. Indeed, it is hard to envisage any place in that process for the written word. ⁴³³  Since oral activity leaves no record, and we are arguing that there is no reason to expect that technical navigational matters would have been described in documents that might have subsequently been lost, that leaves us with virtually no tangible evidence.

Memory and orality must have played a large part in the use of the portolan charts. The interrelationships between the two are complex: what was remembered could be passed on orally, to create a new memory in someone else. Oral exchanges might have been involved even before the voyage began, with input from harbour-masters and others in relation to the pelagic course to be followed at departure and from fishermen for local navigational details. ⁴³⁴ 

Two specific types of memorised information can be distinguished, even if they sometimes merge or reinforce one another:

Narrative – the content of memorised instructions that could be passed on orally. The obvious example would be the directions for a particular course and the distance or time involved

Experience – what was held as visual memory, in the form of the equivalents of a personal photograph album (‘imagebank’) and video records. The ability to recall and recognise coastal features would have complemented the factual narrative of a reimagined voyage

Whoever purchased an early chart could never have encountered anything like it before, even if (which is unlikely) they had previously seen a map of any kind. The chart’s form and cartographic language were original. Some of its conventions might have been assimilated easily, but the symbols denoting dangers for example, which seem to be a portolan-chart invention, were not intuitive. The red dots for shallows could equally have indicated sandbars or mudbanks. In other words, the portolan chart was incomplete, and of much reduced usefulness, without exegesis. They were not designed for unassisted novices. For these reasons, instructions for the charts' use must have been conveyed orally: first from the seller to the sailor when purchasing his first chart, and then from him to his colleagues, including no doubt the ship’s boy as part of his apprenticeship (as described by Benedetto Cotrugli). ⁴³⁵  Additionally, by building up a store of memorised knowledge over repeated voyages, the mariner would be able to attach his remembered experiences to the chart’s littoral toponymic sequence.

Scale: The value of the ‘mile’ represented in the distance scale is never stated on the portolan charts (as it would be on later marine charts). Indeed, although the decimal system was already present on the earliest charts, the users of the Carte PisaneAccepted as the oldest surviving portolan chart, c .1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) would have had to be told that the smallest divisions represented five miles not ten. ⁴³⁷  However, it seems likely that the charts’ users would instead have estimated the duration of a course in terms of how many multiples it was of a few clearly remembered standard distances (perhaps between their home port and others), measured off the chart with their hand. ⁴³⁸ 

Winds: If the eight named winds that were used for the mental compass were indicated at all it was by means of an initial letter placed beside a primary wind. The full names would need to be explained

Dangers: The signs for shoals, rocks, and their various combinations have already been discussed. ⁴³⁹  Because they were unprecedented and no key was provided, they would certainly have required explanation. ⁴⁴⁰  Besides that, the initial, highly simplified symbols seen on the earliest charts were to develop by 1400 into something quite different, thus involving some re-learning. ⁴⁴¹  Even though some of the symbols survived into modern times, we cannot now be sure of the exact meaning that the earliest charts were conveying. Identification also presented problems. Few of the offshore hazards have names alongside and when toponyms were provided they would sometimes be transferred into the main coastal sequence nearby, as if they were land features. Even if the danger was placed with some precision, its specific threat and perhaps its name would need to have been explained.

Coastlines: It might be expected that coastal outlines would have conveyed strict reality, at least to the extent that the map’s scale allowed. However, the portolan charts’ creators had no such concerns. What lay between the headlands was of marginal importance for navigation, so the intervening stretches were usually simplified into regular patterns. The capes themselves would often be drawn in a stylistic manner, sometimes reflecting that chartmaker’s distinctive signature. The chart’s user would have soon become aware that little could be learnt about his ship’s position in terms of recognising particular coastal features, although subtle variations in the depiction of the shoreline might hint at different general types. ⁴⁴² 

Rivers: In a similar way, the portolan chart’s user would come to realise that the great majority of river mouths were depicted in an identical fashion, with only a handful of major outlets enlarged, and the four main deltas in the Mediterranean and Black seas given individual treatment. So, when moving along an unknown shoreline, the charts would do little to aid recognition there either. ⁴⁴³ 

Toponyms: It would be reasonable to assume that this element provided the user with simple, precise information. However, a number of the toponyms had ambiguous or portmanteau meanings, leaving the reference unclear unless there was a prefixed initial to distinguish a cape, bay, port, island etc., of the same name. ⁴⁴⁴  Nevertheless, the charts’ toponymy must have served as a mnemonic device, in the form of labels on which to hang not just referents but also detailed memories, both of instructions received and stored images from past voyages.

Geopolitics: Graphic devices identifying ownership of the more significant ports are first seen on Italian charts of about 1320. Since these town flags include some of the earliest instances of recorded vexillology – which means that they might not have been generally recognised – their significance would need to have been orally explained. ⁴⁴⁵  In addition, those armorial pennants were not a reliable guide to geopolitical reality because the information they conveyed was often out of date. Sailors must have relied instead on verbal exchanges in port if they were to be kept apprised of changes of suzerainty. Likely questions would have been: what reception could be expected if they landed at certain places; were pirates operating in the area and if so how best could they be avoided?

Other illustrative elements: The embellishments introduced on Catalan charts from about 1330 onwards, such as ethnographic illustrations, enthroned rulers and small stylised diagrams for towns, might have needed to be explained but they have no relevance for the earliest surviving charts or, therefore, the origin question.

L.1a.2. Information not provided by the charts but necessary for navigation:

Leaving aside the sea signs listed earlier, ⁴⁴⁶ which could not have been conveyed on a chart, there are some indicators that a portolan chart might have included, for example:

Anchorages: On occasions, a sailor would need to know the location of the nearest protected anchorage to ride out an impending storm. Perhaps the charts’ modest scale made that impracticable but in some cases a small anchor symbol, of the type introduced onto later charts, could have been squeezed in. Given the charts’ practicality and the innovative approach of a number of early chartmakers, there was probably a good reason for this omission.

Approaching harbour: This can be the most challenging part of a voyage but the charts do not provide any pilotage details, such as buoys, lights, leading lines and so on. Unless a local pilot was taken on, or any brief comments there might be in a portolano (Plural portolani and more usually called ‘portolans’ – which can be confused with the charts) these were medieval guides to the coastal places around the Mediterranean, listing the distances between them. They also include open-sea (pelagic) routes were accessible to a literate crewman, they would have had to rely on remembered experience. It would have been possible to include enlarged details of a few of the more hazardous harbour approaches, placed inland – and later marine charts did just that. However, as the situation would often change, with silted channels or shifting mud or sand, the omission of this information from the charts had the benefit of saving them from becoming redundant.

Lighthouses: The absence of those is surprising, given their importance as both warning and location devices ⁴⁴⁷ 

Provisions: Likewise, it might have seemed helpful to have included a code for sources of fresh water and places to re-provision, particularly less obvious ones. Again, shifting circumstances would have rendered some of that information obsolete.

Routes: There must have been preferred routes through the Aegean to the Dardanelles and on to the entrance to the Black Sea, perhaps with different tracks for those entering from the west or east. If so, might a pair of simple lines not have indicated those? Evidently, no doubt for good reasons (variable winds, changing political allegiance?), this was never done.

Wrecks: The charts do not specifically warn of the presence of a wreck, unless any of the black crosses for rocks and reefs could have been interpreted in that way. However, most of those would have already been included by implication in the specific hazard warnings, since it would have probably have been on one of those features that the ship had foundered.

Because a sailor’s training was purely practical, involving instruction, imitation and criticism, it is not surprising that there was no textual guide to that process in the Middle Ages and that details of actual pelagic courses had to come from memory and oral instruction.

An earlier section of this essay had focused on the brain and spatial memory. ⁴⁴⁸  It is worth considering further the capacity of memories that had been developed from infancy. M.R. O’Connor noted that “by six, children show a strong positive relationship between hippocampal volume and their episodic memory – the bigger the volume, the greater the ability to recall details of an event.” ⁴⁴⁹  Medieval pilots would probably have started their life at sea as boys and the voyages they undertook would have been held in memory in the form of a succession of events.

Information stored in the head of a sailor whose safety and even life depended on it, would have been far more secure than any simple diagram or scribbled note. As a result, it was hardly likely that the oral tradition would have been rendered entirely obsolete by either the magnetic compass or the portolan chart. The holistic memory of a single long-distance course might have extended to a number of pages, had the full detail been written down. Such memorised records of voyages would have called upon several of the navigator’s senses as well, with the verbal instructions and personal video record of previous experience being only the most obvious. Turning some of that disparate information into a written pilot guide was not realised until the printed versions of later centuries, when artificial aids had rendered most of the memorised information redundant.

Though possibly not all of it. Perhaps some mariners took heed of Socrates’s warning from the 5th century BCE:

In fact, it [writing] will introduce forgetfulness into the soul of those who learn it: they will not practice using their memory because they will put their trust in writing, which is external and depends on signs that belong to others, instead of trying to remember from the inside, completely on their own. ⁴⁵⁰ 

Which view was restated this century by the cultural ecologist and philosopher, David Abram:

“The alphabet and alphabetic literacy does not cause our human estrangement from the more than human natural surroundings. But it makes it possible in a way that is simply not possible for traditionally oral indigenous people who are so deeply embedded in the particular landscape that they inhabit, and they are practising relationship all the time with … the winds and the weather powers … Alphabetic writing makes possible a forgetting of that larger field of active agencies”. ⁴⁵¹ 

Sailors who were being taught the relevant details of a particular route – and they might have been apprentices, having it beaten into them – would have absorbed a mass of varied information. Among the more important was the knowledge needed to leave harbour safely, before turning to pick up the wind or compass direction required for the first of what might be several distinct legs of that voyage. For each subsequent stage, due note had to be taken of any known hazards that had to be actively avoided on the way to the next waymark, which needed to be recognised and perhaps passed at a specified distance. Gauging the passage of time and estimating distance were essential skills, along with the ability to ‘read’ the wind compass and use the available astronomical aids.

It could be expected that a single individual regularly repeating a particular course might have been able to refine its details and thus, gradually over time, arrive more precisely at his destination. But if this was due less to improvement in the sailing directions and more to enhanced sensory and memorised skill, then it would not necessarily have led to any refinement of the guidance that could be passed on to others. In evolutionary terms it would be irrelevant because it would not have left its mark on the collective memory. That is why this thesis argues that one of the seeds out of which the portolan chart grew would have been the realisation that the individualistic and disjointed pelagic course memories could be made more accurate only through a collaborative process, using a graphic medium to combine them into a single permanent record.

It may be instructive to consider similarities between what was required mentally of a medieval pelagic sailor and the experience of London’s Black Cab drivers, whose brain functions have already been discussed. ⁴⁵²  In each case there was a clear motivation to learning: future personal safety for the ship’s boy and, for the trainee cab driver, the realisation that failure would deny entry to a desirable independent career.

The drivers’ training and daily activity have been closely studied, providing the detailed evidence that is entirely lacking in the medieval case. Just as the cab driver must remember the name of every street on his or her planned route as well as the turnings off it, so the open-sea sailor had to retain a wide range of remembered or learnt details. All would have been absorbed orally, and discussed in a port tavern or on the dockside – the mariners’ bourse or free navigational exchange. Until a few decades ago, the Cabmens’ shelters served a comparable purpose in London. ⁴⁵³ 

Cab drivers: traffic problems (temporary or permanent), accidents, road closures and diversions, alterations to one-way systems, or other reasons to alter a usual route

Sailors: ports now open or closed, new wrecks, changes to the route into port, new customs regulations [all orally communicated in port], piratical activity, new sandbanks or mud bars

Recent research is revealing the reinforcing effect of a well-used memory, finding that:

the more time the London Black Cab drivers spent on the job, the more the hippocampus changes structurally to accommodate the large amount of navigational experience … Drivers who spent more than forty years in a taxi had more developed hippocampi than those just starting out ... Conversely, they found that using a GPS excessively might to lead to atrophy in the hippocampus as a person ages. ⁴⁵⁴

It is surely safe to assume that with both medieval pilots and Black Cab drivers the expansion of their knowledge was never ending.

L.2b. The London Underground Map

What follows may reasonably be considered a digression, but it is offered here as a theoretical analysis to highlight some fundamental cartographic principles and promptings. These unite across many centuries what might otherwise be assumed to be wholly unrelated map types.

The second modern parallel – the London Underground Map – demonstrates remarkable underlying similarities with the portolan charts, not least around the issue of orality. ⁴⁵⁵  Each displays a ground-breaking design placed over a simplified structure, and each requires orally-transferred and memorised information for it to be used to the full. As Catherine Delano-Smith noted: ‘Diagrammatic maps are created specifically to provide an answer for those who know what the question is’. ⁴⁵⁶  We might extend that to a parallel statement: 'A diagrammatic map answers the questions that need answering’.

The creator of the Underground Map, Harry Beck, had previously been an engineering draftsman for the Underground Electric Railways Company of London. ^456a  In 1931, he took the existing, traditional map, and extracted its salient features. He then arbitrarily straightened out the railway lines so that they ran in one of the eight main compass directions (octolinearity), and placed the sequence of stations and interchanges along those lines evenly, with limited regard to the actual length of the intervening distances. Thus he took a normal map and turned it into a highly simplified topological structure, a diagram or cartogram, which massively distorted geographical reality. By doing so he was able to offer the user easy, speedy, and unambiguous comprehension – with, like the portolan charts – almost no textual explanation ⁴⁵⁷ [Powerpoint, Slide 27].

Daily commuters (on crowded trains) need further, more detailed information: which carriage will be nearest the exit at the destination station, and on which side will the doors open. Thus, for the experienced user, just as for the sailor with his portolan chart, a combination of map and memory are required. ⁴⁵⁸  Having lived in London for many years, I would be lost, like everybody else, without the descendant of Beck’s map.

I argue later against an assumed mercantile trigger for the invention of the portolan chart ⁴⁵⁹ and also queried the contention that its origin was dependent on the magnetised needle. ⁴⁶⁰  Instead it is proposed here that the chart’s inception could have happened at any time. For its part, the London Underground Map, rather than being commissioned or linked to any evident social or economic impulse, was a purely personal project, submitted to the Underground management at the suggestion of friends. Indeed, the company rejected it at first as being ‘too revolutionary’. ⁴⁶¹  When it was published in 1933, following a re-submission, it was an instant success. The modest improvements and additions that Beck made to subsequent reissues presumably reflect the suggestions of passengers (as was requested at the first publication); in a similar way, feedback from sailors must have been the prompt for the introduction of a few new features into the portolan charts (and certainly the hundreds of added toponyms).

It is clearly appropriate in many cases for market forces or external events to provide the best explanation of the actions of individuals. But can we not allow the possibility of human agency, of an event explicable only in terms of exceptional individuals, working to their own rhythm? And could that not be applied equally to Beck for his Underground Map and to the unknown sailor(s) who conjured up the portolan chart without the help of any precedent?

Given that the carefully written place-names are such a prominent feature, it is natural to assume [as I had done myself] that the charts must have been designed for literate sailors. Yet that does not necessarily follow, since it seems clear that few of the sailors responsible for guiding a ship at that time would have been able to read, even if some of the other crew members might have been literate. Why would navigators have had need of Latin? And why would the vernacular literacy, which was only just beginning in the 13th century, have dealt with artisanal issues?

Part of the reason that we do not hear the sailors’ stories, and hence have been left in the dark about the early history of the portolan charts, may stem from the fact that most of their users could neither read nor write. Hence the little we can learn about the early charts and their shipboard functions has come from literate passengers not mariners. A number of accounts survive from pilgrims, as well as participants in the crusades, but they offer few insights into navigational practice. ⁴⁶² 

In relation to textual support for historical argument (or indeed the lack of it) there is a useful parallel with the shipbuilders, who would have been close neighbours of the chartmakers in a port. The oldest known treatise on that subject was not written until about 1436 by Michael of Rhodes, and even then it was only the state-run establishments that used paper records at all. ⁴⁶³ 

A further relevant parallel can be found in construction plans. If the cathedral builders, for instance, had made preliminary plans, these have not survived. The well-known sketchbook of Villard de Honnecourt (c.1225-35), includes many architectural drawings but, rather than being building designs, it is now suggested that they “were deliberately simplistic and abstracted to serve as coded mnemonic devices for architects who were initiated into the relevant oral tradition”. ⁴⁶⁴  Describing the situation in England, T.D. Sutton cites the paucity of examples of plans attached to builders’ contracts before 1500 (the earliest is 1380). In a familiar echo of the situation with portolan charts, the two early instances of an actual working plan survived only because their vellum was repurposed in a book binding. ⁴⁶⁵ 

A final example contrasts the largely text-free portolan charts with a noticeably different travellers’ map, the cheaply-printed ‘Rom Weg’ map designed for the use of pilgrims going to Rome for the Holy Year of 1500. This included its own separate explanation sheet. ⁴⁶⁶  Clearly, the expected users in those two cases had different levels of anticipated literacy.

Assertion: Even if some medieval sailors were literate, the inexperienced among them could not have found in the portolani (Singular portolano and more usually called ‘portolans’ – which can be confused with the charts) these were medieval guides to the coastal places around the Mediterranean, listing the distances between them. They also include open-sea (pelagic) routes the knowledge needed for navigation.

Around 1200, few sailors would have been able to read. No doubt there were literate landsmen on board a substantial ship – they would be needed for dealing with mercantile matters, for example – but it is highly unlikely that they would have been responsible for navigation planning or for those decisions that had to be taken at sea.

It is improbable that medieval pilots would have learnt the position of the stars from a book. Or even from a diagram, since, when they were using constellations for navigation, by definition they would have taken advantage of the genuine article above them. Where there might have been a need for written instructions was in the predicting of tides along the Atlantic; though that was not relevant for the Mediterranean coasts. Called ‘Establishment of the port’, it allowed the time of high water to be calculated according to the compass bearing of the full and new moons. ⁴⁶⁷  Yet that information was presented for sailors, not in the form of a text, but as a diagram, a simplified map, or a calendar like those in the Catalan AtlasActually a sequence of six panels, two of which cover the area of the usual portolan chart and two others extend east to make this a world chart, c.1375 (Paris, Bibliothèque nationale de France, Manuscrits, MS. Esp. 30) and later portolan atlases.

It might conceivably be the case that illiteracy could even have offered advantages for a medieval mariner, since mental space that might otherwise have been occupied by remembered texts could be freed up for conserving his experience instead. ⁴⁶⁸ 

It may be a useful exercise to divide up the elements that make up the Carte PisaneAccepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) – our only guide, however approximate, to the likely appearance of its lost antecedents – between those that were squarely aimed at literate users and those which could have been understood by non-lettered sailors.

L.3c.1. Information aimed at literate users

Viewed in that light, the Carte Pisane made only slight demands on its literate users. It includes just three textual elements: toponymy, the names of the eight main winds, and a single hazard warning. The pervasive toponymy is the obvious case apparently demanding reading skills. However, we will consider that, counter-intuitively, in the next, combined category instead. That leaves the repeated warning of a lone offshore danger, a supposed rock to the west of the Ionian Islands, signalled dramatically by the repeated word Guardate [watch out]. It is not worth dwelling on that here but this is clearly written in the same ink and with the distinctive capital G and lower-case d found elsewhere over the chart. In other words, it is not an addition by a user. As it happens, there is no rock near there, nor could there be, because of the depth of water. So, an unexplained anomaly. ⁴⁶⁹ 

The other significant textual element – although found only on some later portolan charts – are discursive historical and geographical legends. Those are generally restricted to Catalan charts from 1330 onwards and, in particular, to the more ornate versions designed for landsmen. ⁴⁷⁰  It is notable that neither there or indeed elsewhere do the charts provide any written guidance about their own use.

Had the early charts included written passages, that would have necessitated choosing a particular language. Would it have been Latin (highly unlikely) or one of the emerging vernacular languages, such as Catalan or an Italian dialect? If so, that could have reduced the charts’ relevance to those besides the chartmakers’ countrymen – since we know they were sold all over the Mediterranean. ⁴⁷¹  In that situation, the Greek sailors in the Aegean would have needed charts that used their own alphabet for the toponymy, yet the earliest known examples of those date from the 17th century. ⁴⁷² 

L.3c.2. Information that could be understood by non-literate users

It may seem provocative to suggest that this category could have included the main textual component of the functioning portolan chart, namely its toponymy. Surely this element would have been accessible only to literate sailors? Kevin Sheehan’s reconstruction of a portolan chart found that the place-names, along with the hydrography, were the most time-consuming parts of the drafting process. ⁴⁷³  Since this meant they would be expensive for copyist and purchaser alike, why would so many chartmakers have taken the trouble to add to the initial place-name list unless those who bought and used the charts were able to appreciate that element?

There is another aspect that may have played a part. The seamen gathered in the Italian quayside taverns would have been speaking in a babel of languages. Faced with these various dialects, it seems that there must have been a tacit agreement that linguistic differences should be submerged in a ‘lingua franca’, which allowed broadly standard toponymic forms to be widely shared. The non-literate sailor could have learnt to recognise the shapes of names and their sequence, or to home in on the more important places that had been helpfully highlighted in red. When that conformity was set aside, as, for example, with the idiosyncratic (sometimes unrecognisable) readings on the 1492 chart of the Portuguese, Jorge de Aguiar, ⁴⁷⁴ the convenience of approximate toponymic standardisation becomes clear.

Which side of the orality/literacy fence should we place the mnemonic island shapes, first introduced – particularly for the Aegean and Dalmatian islands – in the work of the Vescontes Pietro Vesconte (fl.1311-c.30) is the first chartmaker known to us by name, and the first professional. A Genoese, he worked in Venice, along with his presumed relative Perrino. Several of his works survive, documenting in particular the growing shape and toponymy of the British Isles. His chart of 1311 and atlas of 1313 are the oldest to be signed and dated and Angelino DulcetiAngelino Dalorto or Dulceti was a Genoese who worked in Palma, Majorca, leaving us dated charts of 1330 and 1339. His were the first to introduce inland and illustrative details (1311–c.40), and sometimes continued thereafter for centuries? ⁴⁷⁵  Each island or islet remains clearly named on those charts, but it is more likely that the signature shape provided for each entity was intended as a more accessible alternative form of recognition for illiterate sailors, who could associate the outline with its remembered coastal profiles and toponym. That group might have included, for example, literate Greek sailors if they could not understand roman script. In a similar way, the flags signalling ownership of the port concerned could have been recognised by any sailor, as a substitute for a written label.

I take here a different line from Ramon Pujades, who considered that “the nautical chart is not a mute object but rather a map laden with writing” and that “such a valuable instrument of written culture as a nautical chart, which would have been practically useless in the hands of an illiterate…” ⁴⁷⁶  Instead, as has been shown here, almost nothing on a functional chart demanded literacy. The absence of explanatory text, and hence the need for oral instruction instead, meant that navigators just continued what they had been doing previously. The same oral process would have been employed for training young sailors in the use of the charts and for the exchange of information about navigational issues when at sea. From this we can infer that the lack of specific guidance for literate users, or indeed on navigational topics in general, was an intentional decision by the chart’s creators. Given the charts’ multi-lingual usership, and the previous comments about the lack of literacy, it is worth considering the possibility that the chart’s creators might have sought to overcome those difficulties by intentionally producing a ‘non-linguistic’ navigation aid.

What mental faculties were required to use a chart? By focusing on literacy and illiteracy, we are in danger of missing an aspect of potentially greater importance: graphicacy. This has been defined as “the communication of spatial information that cannot be conveyed adequately by verbal or numerical means”, inter alia, citing maps and plans. ⁴⁷⁷  Graphicacy is of direct relevance for both the charts’ creation and the use made of them, and specifically for the understanding of the cartographic conventions it introduced.

If we hope to understand the nascent portolan chart I suggest that we need to move away from treating it, first and foremost, as a cartographical production. This essay asserts that the cartography came later; it did not inform the initial chart’s creation as a graphical diagram or a scaled outline plan. ⁴⁷⁸  The portolan chart’s experiential basis, compared to the theoretical or literary constructs of the time, would have been bewildering for those who did not share the truths it revealed: in other words, all except those pilots with the graphicacy skills to ‘read’ the portolan chart, by interpreting new conventions and symbols, aided by an oral explanation of the chart’s main features.

It was not only pelagic mariners, of course, who used portolan charts. Anybody who coasted along an unknown shoreline or sailed out to sea as a passenger would have been able to gain useful information from a chart. Others could have turned to it because it was the only realistic picture of what was known of the world at the time. They would have used it for its geography, not the hydrography. Non-mariners, such as travellers and pilgrims, must sometimes have asked a sailor to show them the ship’s current position and the planned route on his chart. However, if they looked further they would have failed to find the information they might want about antiquities and the holy sites. ⁴⁹⁵ 

Ramon Pujades has estimated that a basic, functional portolan chart was sold at “approximately two Majorcan or Genoese pounds per chart, which amounted to less than two weeks of a naval officer's average salary”. ⁴⁹⁶  Coincidentally, it might well have taken the practitioner a broadly comparable time to produce a single, purely functional chart. ⁴⁹⁷  But viewed in another way, and considering today’s values, that price could have amounted to what would have been a significant outlay of around 1,000 euros, for a single chart that would need to be replaced once it was too worn or damaged. So, besides being clearly necessary, the chart also had to be looked after carefully.

N.1a. How might the market have been established?

It is fair to assume that those involved in the charts’ creation would have been among the first to request (or draw) a copy for themselves, as well as acting as evangelists among the other mariners they encountered. How would anyone have accepted the portolan charts' outlines and copied them faithfully unless they understood that the charts’ revolutionary geometric framework did indeed mimic reality. That can also explain why the hydrography was treated as sacrosanct by the chart copyists (apart from a few early improvements), while some of the other elements were left to personal choice.

Furthermore, the charts’ evident utility might help answer the crucial question as to how a replication business could have been established in the first place, and then have survived through the decades up to the time of the Carte PisaneAccepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) . For the central point here is that, from very early on, once the charts had become useful for actual navigation, sufficient copies would need to have been produced not only to create but also to maintain a market. In other words, the portolan chart would have had to pay its own way.

Almost continuous production, even if only part-time, or involving no more than a single dynastic family, can be assumed as well. Like any novel device brought to market, the chart draftsmen would have needed repeated orders to give them sufficient motivation to persevere. A critical mass, based on a habit of use, would have been essential for both parties. If those assumptions are accepted as probable, that would provide historians with indirect evidence of the charts’ usefulness – even if that might be considered a circular argument.

Long-distance sailors formed part of an unparalleled information network. Those who embarked regularly on trading voyages must have visited more foreign places and encountered a wider range of people than the members of any other profession. In the Middle Ages, most news from foreign parts travelled by sea, which was fast and convenient. It was those knowledge routes and connections that must have helped both to familiarise sailors with the portolan chart and distribute it, thus accelerating its adoption during the early years. Not only did they become ubiquitous, they must also have been their own ‘pollinators’, spreading fresh information that could be used to update the work of other practitioners. The unforeseen consequences of that same interconnectedness can be seen when the Black Death was carried speedily from Crimea to Europe, and then throughout it (which presumably explains the gap in dated charts between 1339 and 1367).

Suggestions that the creators of the portolan chart must have needed financial and organisational help – reflections of an outlook that perhaps unintentionally denigrates artisanal achievements – can be readily countered. There is no evidence that any individual or institution was sponsoring charts, apart from Pietro Vesconte’s provision of portolan charts to illustrate Marino Sanudo’s Liber Secretorum Fidelium Crucis in the 1320s. Had such support been involved it could presumably have applied only to the single Italian port where the first charts had originated. The evident dispersal to other coastal cities after the Carte Pisane, if not before, weakens further any argument for focused external support at the outset. Since that development belongs to the first chapter in the charts’ recorded history, it cannot directly inform the ‘Origins’ debate. However, as has been said before, information from the first documented period may have analogous relevance when conjecturing about what might have preceded it. That apparent fanning out from the unkown birthplace is discussed in the following sections.

N.1b. Not a single lineage

Whatever superficial similarities there might be between the Carte PisaneAccepted as the oldest surviving portolan chart, c.1270, its author and place of origin are unknown. It has been in the Bibliothèque nationale de France since 1839 (Cartes et Plans, B 1118) and the charts that followed, now preserved in CortonaBased on its archaic appearance, this is considered to be the second oldest surviving portolan chart after the Carte Pisane, dating from the beginning of the 14th century or even earlier. It was first described in 1957 (Biblioteca dell'Accademia Etrusca di Cortona, Port. 105), AvignonA chart fragment focusing on the North Sea, probably very early 14th century, discovered in 2002 in the Vaucluse Archives (Port 01, 3E 54) – for Jacques Mille’s studies on this see the Portolan Chart Bibliography: https://www.maphistory.info/portolanref.html and LuccaDiscovered in 2000 in the city archives, this is visually very different from other charts from the 1st quarter of the 14th century (Archivio di Stato, Lucca, Fragmenta Codicum, Sala 40, Cornice 194/1) , as well as the Riccardiana chart The oldest Genoese chart, probably dating from the second decade of the 14th century (Biblioteca Riccardiana, Florence, 3827), Carignano map/chartDrawn by the Genoese priest Giovanni da Carignano c.1325-30, this is a hybrid of a map and marine chart. Florence, Archivio di Stato, C.N. 2 [destroyed in World War II] and Library of Congress chartThe oldest portolan chart preserved in the USA, probably dating from the second quarter of the 14th century, it was acquired by the Library of Congress in 1914 – which should presumably be arranged (in what order?) in the period between the Carte Pisane and the end of the Vescontian output (c.1270-c.1330) – there are no pointers to a single centre for portolan-chart production over that period. While the last three of those charts are evidently of Genoese origin, there seems to be no good reason why the others could not have been made in four different Italian ports. ^497a 

It is also worth pointing out that three out of that group of seven have unique elements, which suggests they might have been created for other purposes than navigation (not being used at sea may also explain their survival). As described by Jacques Mille, who has studied it very carefully, the Avignon chart is the earliest survivor to treat the North Sea coasts, and, uniquely, includes profiles of ecclesiastical buildings; ⁴⁹⁸ the Lucca chart has elaborate city signs; ⁴⁹⁹ and the Carignano map mixes the characteristics of both a marine chart and a topographical map. ⁵⁰⁰  Nevertheless, among that list of early unsigned charts there are some shared features. We are left to speculate about the possible significance of the supplementary squared-off compass-line grid on the Carte Pisane and the Avignon chart (and the equally incomplete system on the Cortona chart), the repeated guardate, guardate (indicating a non-existent rock) on the Carte Pisane and Lucca charts, or the partially overlapping toponymy of the Carte Pisane and Cortona chart. Those occasional shared features are outweighed, though, by the disparities within those pairings. The overall impression is one of stylistic heterogeneity; nor is there any justification for assuming that those early survivors are in any way representative of what was being produced during that period.

Each chartmaker would have copied the coastlines and offshore danger markings from an earlier exemplar, while simultaneously reproducing most of the same toponyms. In other words, they shared a common cartographic ‘language’, which indeed is what provides the best (if unscientific) definition we have for a portolan chart. Given that the chartmakers were mostly port-bound, how could any consistency in the charts’ content have been maintained? The likely answer is that each was kept informed of developments elsewhere through being shown the charts of other makers.

Yet there was little borrowing of the more obvious visual features. Chartmakers, it seems, wanted their work to be recognisably distinct from their forebears and competitors, a tendency that can be seen continuing into the later centuries.

Had any two of those seven very early charts been in the same hand, we could probably have determined which was the earlier of the two, thus helping to identify at least one practitioner’s development, but that is not the case. From another direction, when trying to order the seven chronologically, there is a valuable tool in the largely reliable yardstick of the expanding toponymy and the emerging outlines of the British Isles introduced into Pietro and Perrino Vesconte’s Pietro Vesconte (fl.1311-c.30) is the first chartmaker known to us by name, and the first professional. A Genoese, he worked in Venice, along with his presumed relative Perrino. Several of his works survive, documenting in particular the growing shape and toponymy of the British Isles. His chart of 1311 and atlas of 1313 are the oldest to be signed and dated charts in the 1310s and 20s – on the assumption that those additions were indeed introduced by the Vescontes and not somebody else from whom he had, in turn, borrowed. However, that does not allow us to say confidently that a particular chart was drawn before or after Vesconte’s productions of, say, 1318 because there is rarely a precise match. The Riccardiana chart, for example, can be placed earlier or later than Vesconte’s atlas of 1313 depending whether the focus is on its toponymy or the Atlantic outlines. ⁵⁰¹ 

Counter-intuitively, therefore, it seems that the concentration of portolan-chart production in organised workshops in (mostly) three cities, Genoa, Palma (Majorca) and Venice, from perhaps 1320 until well into the 15th century, had been preceded by a period of distributed, informal local production in family ateliers, ⁵⁰² and would later be followed by a dispersal among a growing number of major Mediterranean ports.

Prior to the circulation of printed texts in the later 15th century, the portolan chart must have been reproduced (sometimes with intentional modifications) in greater numbers, not just than any other map but probably any other coeval document besides religious texts. ⁵⁰³ 

The oldest direct reference to what must have been a marine chart is 1270; ⁵⁰⁴  they are mentioned a number of times in the 14th century and, by the 15th (if not earlier) they were being talked about as essential aids. ⁵⁰⁵  Hence, if a Mediterranean mariner had encountered one, say in 1220, it would probably have been the first map he had seen. By 1320 they must have been circulating generally and, well before 1420, they would have become commonplace.

The 180 or so charts and atlases that survive from before 1500 were joined by many hundreds in the 16th and 17th centuries. ⁵⁰⁶  It also seem likely that a wholesale trade would have grown up, perhaps before it was first documented in the 15th century. ⁵⁰⁷  The 24 charts that Gabriel de Vallseca was required to provide for a merchant to sell on, gives us a glimpse of that process. ⁵⁰⁸ 

Because of the exigencies of use at sea, each surviving functional chart might represent hundreds, or more likely thousands, of those actually produced. ⁵⁰⁹  Essential, and also ubiquitous, the portolan chart would have been found on nearly all long-distance trading vessels and could probably have been purchased in most Mediterranean ports. No Christian mariner could have been unaware of them, while Islamic sailors clearly used them and made copies in their own script.

The question of apprenticeship – how it might have operated and when it was perhaps introduced – follows naturally from the preceding section. Perrino Vesconte (fl. 1321), presumably the son or nephew of Pietro (fl. 1311), the first chartmaker to leave us a signature, seems to have been trained to produce work that was indistinguishable from Pietro’s. That is what we would expect from an apprentice, who would usually have started out when very young. The training process would probably have included being taught to read. The apprentice would also have been required to insert the toponyms in imitation of the master’s hand, and reproduce very precisely the house-style for each of the chart’s elements. Nevertheless, if that Vesconte instance indicates some kind of apprenticeship, it is exceptional.

Despite assiduous searching, Ramon Pujades found no apprenticeship records for Catalan workshops in Palma, Majorca, before 1368 and, for Italy, before 1427. ⁵¹⁰  There may well have been other instances of which we are not aware but no signs of apprenticeship have been identified among the early surviving charts. Instead, what is immediately evident are the obvious differences, as was found when comparing, for example, those six early unsigned charts: Carte Pisane, Cortona, Avignon, Riccardiana (Florence), Lucca and Library of Congress. ⁵¹¹  Had there been apprenticeship – in other words if those who drafted the charts had learnt their craft during their formative years – we would expect to see close imitation in both the charts’ fundamental structures and their overall style.

One specific feature that is worth a full study is the handling of the twin-circle compass network, which was prevalent up to the end of the 14th century. Where the circles meet (at the point of tangency) the area above and below falls outside the network of compass lines, so that anyone sailing in those areas would have had to mentally supply the missing directions. The various methods devised to deal with that omission, which affected what was the most routine of the chart’s elements and the one that was probably inked in first of all, are likely indicators of a chartmaker’s stylistic signature. For example, the Carte Pisane and the Avignon chart bridge those gaps, rather clumsily, with sections of squared grids. Unfortunately for this diagnostic purpose, the limited coverage of two of the other four charts mentioned above ruled out the need for twin circles. Hence, only those in Florence and Lucca had to deal with that problem, which they manage more confidently with additional, but different, geometry.

What conclusions about apprenticeship might we draw from this? There are two likely ways a copyist could have obtained the complete chart he needed for his model: (a) from copying the pattern used in the atelier where he was working, and then setting up on his own, or (b) simply transferring the details from an existing chart. The first method is likely to have reproduced both the constructional method used for the model and its detailed content; but in the second instance – where the copyist had not been trained in that particular style – new procedures would have had to be devised. The tight discipline enforced on a trainee copyist is likely to reveal itself through the repetitions of distinctive features in his own later productions, which can serve to identify the master-apprentice link ⁵¹² [Powerpoint, Slide 28].

Had the nascent chartmaking production been concentrated in a single port – and Genoa is sometimes mentioned in that context without, in my view, convincing evidence ⁵¹³ – we would expect any expanding ‘school’ to have developed its own shared characteristics, as would later be the case in Majorca. But there is no sign of that in the early decades.

Each chart was of course hand-copied, but that does not mean they had to be identical. Although this might only apply once a market had been fully established, the purely functional examples could well have been pre-prepared so as to be available for immediate sale. The date and signature (if that was thought necessary on such a mundane object) might have been added at the time of purchase. ⁵¹⁴  One or two of the examples that included a limited amount of illustration might also have been drawn beforehand for off-the-shelf sale. But it seems likely that those charts with a considerable amount of adornment would have been made, over a period of weeks or even months, to a customer’s specific requirements. ⁵¹⁵