Saussure and his intellectual environment

SUMMARY The present study paints the intellectual environment in which Ferdinand de Saussure developed his ideas about language and linguistics during the fin de siècle. It sketches his dissatisfaction with that environment to the extent that it touched on linguistics, and shows the new course he was trying to steer on the basis of ideas that seemed to open new and exciting perspectives, even though they were still vaguely defined. As Saussure himself was extremely reticent about his sources and intellectual pedigree, his stance in the lively European cultural context in which he lived can only be established through textual critique and conjecture. On this basis, it is concluded that Saussure, though relatively uninformed about its historical roots, essentially aimed at integrating the rationalist tradition current in the sciences in his day into a new, ‘scientific’ general theory of language. In this, he was heavily indebted to a few predecessors, such as the French philosopher-psychologist Victor Egger, and particularly to the French psychologist, historian and philosopher Hippolyte Taine, who was a major cultural influence in nineteenth-century France, though now largely forgotten. The present study thus supports Hans Aarsleff's analysis, where, for the first time, Taine's influence is emphasised, and rejects John Joseph's contention that Taine had no influence and that, instead, Saussure was influenced mainly by the romanticist Adolphe Pictet. Saussure abhorred Pictet's method of etymologising, which predated the Young Grammarian school, central to Saussure's linguistic education. The issue has implications for the positioning of Saussure in the history of linguistics. Is he part of the non-analytical, romanticist and experience-based European strand of thought that is found in art and postmodernist philosophy and is sometimes called structuralism, or is he a representative of the short-lived European branch of specifically linguistic structuralism, which was rationalist in outlook, more science-oriented and more formalist, but lost out to American structuralism? The latter seems to be the case, though phenomenology, postmodernism and art have lately claimed Saussure as an icon.

4.2. The case of Egger 4.3. The case of Pictet 5. Epilogue on methods of argumentation

Introduction
Ferdinand de Saussure , though a universally known figure in and outside modern linguistics, has remained an enigmatic source of controversy and confusion ever since the posthumous publication, in 1916, of his now world-famous Cours de linguistique générale (henceforth Cours), published by Payot in Paris/Lausanne. This book was, though published under his name, not written by him but based on lecture notes taken by his students (and one or two other, minor, sources) and collated by Charles Bally  and Albert Sechehaye , with the help of the student Albert Riedlinger, who had actually followed Saussure's lectures. Bally and Sechehaye were both Swiss, like Saussure himself, and both already established linguists at the time of Saussure's death. Bally was Saussure's successor in the Geneva chair of general linguistics and comparative Indo-European studies. Sechehaye had been Saussure's student during the 1890s and succeeded Bally in 1939. 1 The fact that Saussure did not himself write the text of his Cours is important, though too often forgotten. His two editors, Bally and Sechehaye, apologise profusely in their Préface to the first edition, saying that they had to try to 'arrive at the thought that we only had the echoes of' (8) and that '[t]he maître would perhaps not have authorised the publication of these pages' (11). 2 Godel states that Saussure was definitely averse to giving these lecture courses, as he felt depressed, uncertain and inadequate: Towards the end of 1906, he was appointed to offer a course in general linguistics at the University of Geneva, where he had been teaching Sanskrit and comparative philology for fifteen years. A friend of his told me that this new appointment simply terrified him: he did not feel up to the task, and had no desire to wrestle with the problems once more. However, he undertook what he believed to be his duty. 3 One may infer that Saussure, had he still been alive, would indeed not have approved of the publication of the Cours. It should, therefore, be borne in mind throughout the present study that when we speak of Saussure in the context of the Cours, what is meant is the reconstitution of his words and views as we know them through the text of the Cours. Standards of authorship thus apply only partially. At the same time it must be realised that it was the book, not Saussure's own words, thoughts or notes which became better known to the world at large in the course of the twentieth century, that has become so influential, first in linguistics proper, then in postmodernist philosophy and art critique, despite its unclarity and the controversy around it.
The lectures in question consisted of three semester courses given in 1906-07, 1908-09 and 1910-11 at the University of Geneva. The Cours as published in 1916 does not distinguish the three courses from each other, as the editors decided to amalgamate the students' notes they had at their disposal from all three courses into one single text. Again, as the Cours has had its influence as a single book, not as three successive courses, it would seem that nothing much hangs by such a historical distinction. It would be different if it could be shown that a major change occurred in Saussure's thinking during those five years, but the detailed material that is still extant and has meanwhile been made 1 In 1908 Sechehaye published a book Programme et méthodes de la linguistique théorique: psychologie du langage (Paris, Champion; Leipzig, Harrassowitz; Geneva, Eggimann), which contains ideas, especially on synchrony and diachrony, also found in the Cours. Circumstantial evidence (see Peter Wunderli, 'Saussure als Schüler Sechehayes?', in Studies in the History of the Language Sciences, 9, 'In Memoriam Friedrich Diez', edited by Hans-Josef Niederehe and Harald Haarmann (Amsterdam, 1976), 419-60, reprinted in Peter Wunderli, Saussure-Studien: exegetische und wissenschaftsgeschichtliche Untersuchungen zum Werk von F. de Saussure (Tübingen, 1981), 180-200, and Peter Wunderli, '"Acte", "Activité" und "Aktion" bei Saussure', Cahiers Ferdinand de Saussure, 42 (1988), 175-201 suggests that, as regards those parts of the Cours that deal with this distinction, Saussure was, in fact, to a considerable extent indebted to Sechehaye, whom he never mentions. public contains no indication that this happened. There are plenty of testimonies that Saussure was in a perennial state of self-doubt and insecurity about the new ideas he propounded during the three lecture courses, leading to a recurring pattern of intellectual hesitation and fluctuation, but that pattern remained the same throughout. Saussure had not always been so insecure. On the contrary, as a young man he had been both cocky and intellectually precocious. At the age of eighteen he became a member of the Société de Linguistique de Paris. He soon sent in six papers to be published in the Société's Mémoires of 1877, all on single etymologies and other specific matters of comparative-historical linguistics. 4 At the age of twenty-one he had produced a book of 326 pages, his Mémoire sur le système primitif des voyelles dans les langues indo-européennes of 1879, 5 in which he proposed a daring and ingenious hypothesis about the original Indo-European vowel system, which was dramatically confirmed by the Polish linguist Jerzy Kuryłowicz, 6 after the decipherment of Hittite and its recognition as an Indo-European language in 1915 by the Czech linguist Bedřich Hrozn. This alone would have been sufficient to secure him a prominent position in the history of linguistics. Immediately after the Mémoire, he wrote his PhD thesis for Leipzig University on the Sanskrit genitive absolute construction. After these precocious achievements, however, there was a dramatic decrease in the volume and nature of his publications. Until his death in 1913 he published a modest number of mostly very short articles, all on topics of etymological or, occasionally, dialectological detail, and a few book reviews. On the ideas expounded in the Cours he never published a word, although it appears that he started developing these ideas from 1880 onward, the year in which he moved to Paris, where he taught Gothic and Old High German at the École Pratique des Hautes Études. He stayed in Paris, on and off, until 1891, when he moved back definitively to his hometown of Geneva to take up a chair in comparative-historical linguistics that had been offered to him.
This story suggests that Saussure was a highly competent Indo-Europeanist, who had shown flashes of genius during his younger years but lost his brilliance once his youth was over. The man who gave the lectures published in the Cours is very different from the brilliant youngster, no longer the flamboyant young scholar breaching intellectual barriers in comparative linguistics, but an already ageing troubled soul who tried hard but ultimately felt incapable of breaching the barriers blocking a deeper insight into the nature of language. In fact, such was his state of mind that he seriously engaged in occult speculations about hidden messages 'from beyond' in Greek and Latin poetry, based on occurrences or repetitions of syllables and sounds. 7 This aspect is left out of consideration in the present study.
From the 1870s on, linguistic theory has been characterised by the drive to turn linguistics into a 'real' science, first in a comparative-historical or 'diachronic', then in a 'synchronic' sense. Saussure typically straddled the transition from the earlier to the later stage, and it was in this context that, after a difficult start, 8 his reputation as the initiator of a 'new' linguistics, rapidly started growing during the late 1920s, eventually to take on the well-nigh mythical proportions it has today. 4 John Joseph, Saussure (Oxford, 2012). 5 Ferdinand de Saussure, Mémoire sur le système primitif des voyelles dans les langues indo-européennes (Leipzig, 1879). 6 Jerzy Kuryłowicz, 'ǝ indo-européen et ḫ hittite', Symbolae grammaticae in honorem J. Rozwadowski (Krakow, 1927), I, 95-104. 7 Jean Starobinski, Les mots sous les mots: les anagrammes de Ferdinand de Saussure (Paris, 1971). (English translation by Olivia Emmet: Words upon Words: The Anagrams of Ferdinand de Saussure [New Haven, 1979]). 8 Saussure's reputation as an innovator of linguistic theory took some time to get established. In 1924 the American linguist Leonard Bloomfield, the main founder of American structuralism, wrote a polite but disparaging review of the second (1922) edition of the Cours in the Modern Language Journal, criticising Saussure for taking the word and not the sentence as the primary unit for linguistic analysis and description (Leonard Bloomfield, Review of Cours de linguistique générale by Ferdinand de Saussure, Modern Language Journal, 8 (1924), 317-19). In 1926 the Dutch linguist Jac. van Ginneken wrote, in German, an acerbic note of three lines in Indogermanisches Jahrbuch about the same second edition (Jac. van Ginneken, Notice on F. de Saussure, Cours de linguistique générale, 2nd. ed. 1922, Indogermanisches Jahrbuch, 10, no. 5 (1926, 37): 'Unchanged reprint of the first edition [not quite correct, as shown below in the text; PAMS]. Of very uneven value. Dilettante extravagancies, next to deep insights into the life of languages.' ('Unveränderter Abdruck der ersten Ausgabe. Sehr ungleich von Wert. Neben tiefen Einsichten in das Leben der Sprache, dilettantische Freisinnigkeiten.') The more positive reception came later.
Paradoxically, however, what is now known as European structuralism in a wider sense, though said to be founded on Saussure's Cours, became the opposite of 'scientific' in any accepted sense of the term. At present, Saussure's reputation is greatest among scholars of art and literature, among postmodern, mostly French, philosophers and among those linguists who have a predilection for the kind of theorising that is judged correct or incorrect according to whether or not it is felt to resonate with intuitive, introspective or visceral feelings and experiences-scholars, in other words, who abhor formalised theories of the kind found in the more exact sciences and meant to predict empirical data as touchstones. 9 The paradox is solved, at least in part, if one takes into account that Saussure had no knowledge of, or perhaps was repelled by (we don't know), the (neo)positivist wave of formalisation, mathematicisation and data-sensitive empiricism in the sciences that started in the late nineteenth century and has largely defined the notion of science in the twentieth. His idea of 'science' and 'scientific' predated this modernist movement. For him, 'scientific' implied complete classifications and intuitive understanding of 'facts' that really should be of a physical nature but alas, in the case of language, were of a mental nature. More positivist-inclined American structural linguists, such as Leonard Bloomfield and the whole behaviourist school of American structuralism, proceeded by banishing, admittedly by sleight-of-hand, the realm of the mental from 'science', considering it inaccessible to scientific methods. This 'sleight-of-hand' was exposed by the founders of cognitive science, which came into being during the 1960s but has so far likewise been unable to bridge the gap. Yet this structuralism, which was behaviourist at first but became 'mentalist' after 1960, 10 turned out to be exceedingly successful at uncovering rule systems, placing the facts of language in a new explanatory light and opening up entirely new perspectives. Those who now revere Saussure are basically averse to these new developments in the study of language and prefer to stick to a notion of 'science' that is a great deal less rigorous and in fact antiquated and considered unacceptable in present-day philosophy of science.

Saussure's problem with his intellectual environment
One disturbing feature of the Cours, and of Saussure generally, has contributed considerably to the general lack of clarity regarding Saussure's views and his position in the history of linguistics: the fact that he hardly ever mentions any contemporary or older authors that he either follows or opposes. According to Joseph, this was due to the character of the Cours, which was not 'prepared [ … ] for print'; 11 'He was after all teaching courses, not writing a book.' 12 But surely, mentioning relevant 9 There has been a recent spate of publications, (e.g. John Joseph, 'Pictet's Du beau (1856) (New York, 2015), attempting to restyle Saussure as being rooted in romanticist, not rationalist, thought. Joseph is discussed below in detail. Gasparov is not further discussed, as it shows a lack of familiarity with basic historical facts, both in and outside linguistics. Stawarska argues that the Cours does not reflect the real Saussure at all and that the editors Bally and Sechehaye willfully misrepresented the lectures as they were really given. For this author, the real, or 'private', Saussure was a deep philosophical thinker, who was a phenomenologist at heart. She presents Saussure as the founder of a 'phenomenological linguistics', which, however, is left undefined. Unfortunately, this book is marked by absence of proper academic argument: the 'arguments' presented are so biased and selective that one can only speak of rhetoric and propaganda. Then, Bally and Sechehaye, the editors of the Cours, are (unjustly) debunked as the creators of a strawman, the 'public' Saussure known from the Cours, but the fact that it was this 'strawman' that helped shape European structuralism in whatever sense of the term, while the private Saussure had no influence at all, remains in limbo. Finally, even if the private Saussure were indeed a closet phenomenologist, quod non, the relevance of Stawarska's claims is unclearunless it is assumed that this author, who, like Gasparov (2012), is signally unfamiliar with linguistic theory and practice, aims at promoting a new paradigm in linguistics, her 'phenomenological linguistics', for which a restyled 'phenomenological' Saussure is meant to serve as an emblem.
authors is a requirement not only in academic publications but also, and equally, in serious academic teaching. Moreover, this ignoring of other authors was already evident in his Mémoire of 1879. Yet, as we know from Joseph's Saussure monograph, he was extremely anxious that others should not purloin his own ideas without attribution. Apparently, Saussure had a problem finding his place in his intellectual environment. Not mentioning relevant work by others got him into trouble already with his Mémoire in 1879. Ambitious young, but already established, scholars such as Karl Brugmann  or Hermann Osthoff (1847-1909, who had been teaching Saussure in Leipzig, found many of their own ideas in the Mémoire without attribution and felt slighted to the point of privately accusing him of plagiarism. 13 No charges were brought but the taint stuck: the Mémoire was mostly ignored in Germany, though praised to excess in France. 14 It does not look as if the palpable chauvinism that existed on both sides suffices by itself to explain this difference in reception.
In the Cours, Saussure does not enter into a dialogue with existing literature and does not define his position vis-à-vis prevailing ideas, thereby-no doubt unwittingly-creating an impression of greater originality than was warranted. There is no reason to assume malice. It is much more likely that he was so engrossed in the development of his own ideas that he could no longer see to what extent he was indebted to others, or did not bother about it-a mild form of autism that comes naturally with creative enterprises. Tracing the intellectual influences he was no doubt subject to, and also spotting the strains of thought that he rejected and did not wish to be associated with, thus becomes a matter of conjecture and textual critique.
Joseph's Saussure of 2012, a massive volume of 780 pages, is an extensive biography of Saussure, driven by admiration for the man and detailing his aristocratic family history, his schooldays, his early attempts at poetry, and following him, whenever possible, from day to day throughout his later years. As such it is a very useful book, but what is missing is a positioning of Saussure in the wider historical and cultural context of his day. One is not told where he stood, for example, with regard to the great currents of the enlightenment and romanticism, such as the Cartesian-Humean-Kantian problem of the justification, reliability and objectivity of knowledge, which dominated all European philosophy from René Descartes down to Bertrand Russell and was classically treated by Immanuel Kant, or with regard to the great debates, raging all over, concerning positivism and scientific method, 15 or the relation between matter and mind, or body and soul, likewise first broached in its modern form by Descartes but strongly present during Saussure's lifetime, or the ancient problem of the relation between the mind, logic and language, in particular the great subject-predicate debate that had started around 1850 and was central to linguistic theorising during his lifetime. 16 Joseph's silence on these matters would suggest that Saussure was not much interested in, or au fait with, these larger cultural and historical issues, even though they were very much at the centre of debate among the intellectuals of his day. Saussure seems to have had a greater natural affinity with specific questions of etymology than with the wider landscape of overarching philosophical and theoretical issues in the humanities, which seem to have largely passed him by. His knowledge of history was also rather limited, if we may go by the inadequate and chauvinistic Chapter 1 of the Cours, 'Coup d'oeil sur l'histoire de la 13 Joseph, Saussure, 244-45. 14 Joseph,Saussure, See Donald G. Charlton, Positivist Thought in France during the Second Empire 1852-1870(Oxford, 1959, who speaks (11) of the French Second Empire as 'a time when ideas are "in the melting pot", a time of dissension and confusion'. That Saussure, as far as is known, never ever referred to the then universally debated question of the status of subject and predicate in natural language sentences is as baffling as it is damning. Yet it has, to my knowledge, never been noted in circles of Saussurologists.
linguistique' ('Brief survey of the history of linguistics'). In any case, if Saussure was really not, or hardly, in touch with the great philosophical and cultural issues of his day, this must have formed a serious handicap in his attempts at a general theory of language, a new theoretical linguistics. This, one surmises, is why he relied, for such matters, on the 'higher authority' of men like Hippolyte Taine or Victor Egger.
Perhaps due to the tradition in his family, which counted some eminent scientists among its members, Saussure's affinity lay more with the natural sciences than with the humanities. One of the things that attracted him in the new school of the Young Grammarians, 17 which suddenly arose, as if by explosion, at Leipzig during the late 1870s and whose members taught him there, was the fact that these enterprising young linguists had decided to study language as naturalists, that is, in its natural ecological and social environment. Each specific language was seen as a social institution deposited in the minds of its speakers, and no longer as a reified 'organism', a disembodied separate sphere of reality or, 'a fourth realm of nature' (Cours, 17), next to the physical, the mental and the social, beyond the methods of established science, as had been done by earlier comparative linguists. Yet 'no matter how great the services rendered by this school [i.e., the Young Grammarians; PAMS], one cannot say that it has shed light on the whole complex of questions; today the basic problems of general linguistics are still awaiting a solution' (Cours,19). It was precisely these 'basic problems', whatever they might be, that he aimed at addressing in the Cours in what he saw as a scientific way.
It is unclear, however, how far his knowledge of the theory and general principles of science actually went. He never studied any science or philosophy of science, but he grew up in a strongly science-oriented family-a fact that seems to have limited rather than expanded his horizon. His famous great-grandfather Horace-Bénédict de Saussure (1740-99), who, according to Joseph, 18 set the intellectual standard for the family for generations to come, was not a theoretician but rather a naturalist, explorer and inventor, practising geology, botany, chemistry and the like, especially with regard to the Alpine mountains near Geneva. Saussure's grandfather Nicolas-Théodore de Saussure (1767-1845) was, though perhaps less famous than his great-grandfather, a notable chemist and plant-physiologist. His father, Henri Louis Frédéric de Saussure , was an adventurous but respected mineralogist and entomologist. His younger brother René de Saussure (1868-1943, with whom Ferdinand was in close contact, was a mathematician and promoter of the international auxiliary languages Esperanto and Interlingua. But the philosophy and methodology of science were not in the family book.
He knew of, but apparently very little about, traditional logic, though his metaphorical use of the term 'algebra' in the Cours may suggest that he had heard, perhaps through his brother René, of the new algebraic approach to mathematics and logic initiated by George Boole (1815-64) in Ireland. Yet there is no evidence that he was familiar with Boolean algebra or its later application in classical set theory. The second wave of formalisation of logic and mathematics, initiated during the 1880s by men like Gottlob Frege (1848-1925) and Giuseppe Peano (1858-1932, and given a canonical foundation by Alfred Whitehead (1861-1947) and Bertrand Russell (1872-1970 in their monumental Principia Mathematica, 19 seems to have passed him by completely. In this respect, Saussure was a child of his immediate environment, which, on the whole, took little notice of these new developments.

17
I use the term Young Grammarians for what are commonly known as Neogrammarians (French Néogrammairiens). The term Neogrammarians is, though in general use, a misnomer, resulting from ignorance of the fact that the original German moniker Junggrammatiker, coined in 1878, was a jocular takeoff on the names used in those days for angry young nationalists, such as the Jungdeutschen, Junghellenen or, especially, Jungtürken, called Young Germans, Young Hellenes and Young Turks in English. See Seuren, Western Linguistics, 92-93, for the full story. The main feature of the YoungGrammarian doctrine was the 'absolute regularity of sound change', a principle to which Saussure explicitly and uncritically adhered (e.g. Cours, 198). 18 Joseph, Saussure, 16. 19 Alfred N. Whitehead and Bertrand Russell, Principia Mathematica, 3 vols. (Cambridge, 1910(Cambridge, -1913.
What is known about Saussure is consistent with the notion that his intellectual and historical outlook was rather limited and that his ideas developed largely in the context of the common deposit accumulated in the French intellectual world of his day, in which he defined his own position and to which he made his own contribution. This would mean that he, though tuned to the intellectual landscape he lived in, was relatively uninformed about the ways it had come about. If that is so, then understanding Saussure requires a historical analysis of his intellectual environment. Such an analysis then reveals that, in principle, Saussure joined the ranks of the rationalists, not the romanticists, and that his chief aim was to integrate the old rationalism, about which he knew too little, into the modernist strands of thought current in his day, where, again, he remained a relative stranger, and to apply the results to a new, 'scientific', general theory of language.
In the rationalist logico-philosophical tradition of language studies, all the way down to Aristotle, a prime concern had been the triadic relation among: (i) the concepts and logical structure of propositions; (ii) the lexical meanings plus the syntactic structure of natural language sentences; and (iii) the structure of reality. This tradition, however, came to a sudden end during the seventeenth century, only to re-emerge around 1850, with humps and bumps, in the works of Sir William Hamilton (1788-1856) (unjustly dismissed by modern mathematical logicians) and Augustus de Morgan (1806-71), followed by Frege, Russell and the entire twentieth-century development of model theory and formal semantics, and in general the formal study of language.
The Grammaire générale et raisonnée, published in 1660 by the Port Royal members Claude Lancelot and Antoine Arnauld and heavily indebted to the Spanish linguist Franciscus Sanctius or Sánchez (1523-1600), the last great innovator in linguistic theory before 1600, was the farewell call of the old rationalism in linguistics. It owes its fame and influence during the subsequent centuries mainly to the fact that it was the last available source for a tradition that lay dormant under the weight of enlightenment and romanticism but was nevertheless indispensable to serious grammatical work. 20 With the exception of the Port Royal Grammar of 1660, whatever publication on the formal relations between language, thought and world may have seen the light of day during the period between, say, 1650 and 1850, failed to have any impact, as all attention went to the more ecological, psychological and experiential aspects of language and its various uses.
While, of course, the exploration of these new psychological and social dimensions has been of enormous importance to our present-day insights into the nature of human language and of humanity as a whole, it must be recognised that the old rationalist tradition, especially as it re-emerged after 1850, has produced results of similar magnitude and importance. It is a historical tragedy that, over the past century and a half, the protagonists of the two currents of thought have, with very few exceptions (such as the Oxford school of Ordinary Language Philosophy, which was active between 1945 and 1970), steadfastly refused to try and forge a synthesis that would unite the two in a rational manner. Today's world of linguistics suffers heavily from that failure.
On the whole, one may surmise that the urge felt by Saussure to construct the framework for a new 'scientific' theory of language found its origin in his family-bred affinity with the natural sciences: his aim was to give linguistics a new status as a natural science or at least a discipline as 'scientific' as the work done by his great-grandfather, the famous naturalist, and other members of his family. His beloved etymologies and the sound laws formulated by the Young Grammarians who taught him in his early years at Leipzig footed the bill only in part. He felt that more was needed, and in this he was not alone.
But here again, in the narrower circle of linguistic theorists, Saussure seems to have suffered from insufficient interaction with his colleagues. The only two linguistic theorists mentioned more than once are Saussure's fellow-Genevan Adolphe Pictet, about whom more below, and the American linguist William Dwight Whitney (1827-94), who was likewise trying to create a more 'scientific' foundation for the theory of language but whose work, in actual fact, showed little overlap with Saussure's. Saussure had met Whitney in Berlin in 1879 while preparing his PhD, and greatly 20 For extensive comment, see Seuren, Western Linguistics, 46-48, 79, 105-6, 464; Seuren, 'The Logico-Philosophical Tradition'. admired the elder scholar. 21 Apart from the handful of names in the brief historical survey of Chapter 1 mentioned above, Whitney is one of the very few authors actually named in the Cours, on p. 26 and on p. 110, where he is praised for his views on the social character of languages and for his notion of the arbitrariness of the linguistic sign, though also criticised for not going far enough and not seeing that (Cours, 110) 'this arbitrary character radically separates language from all other institutions' (which, by the way, is not quite correct, as rituals are also often either arbitrary or based on historical accident).
Saussure is silent, for example, on Friedrich Max Müller (1823-1900), the famous German-born Oxford linguist, indologist and founder of comparative religious studies, who also wished to see linguistics as a science, but in a different way from Whitney. While Whitney saw linguistics mostly as a (socio-)historical science, Müller envisaged a scientific linguistics in which language is like, or perhaps even actually is, a biological organ that grows and evolves in its ecological environment the way biological organs do. Whitney and Müller were engaged in a bitter polemic, known all over the world of linguistics. But neither Müller nor the controversy is ever even alluded to by Saussure, even though Müller's views impinge directly on those Saussure was developing.
Nor is the German-Austrian linguist Hugo Schuchardt (1842-1927) ever mentioned in the Cours, though Schuchardt too strove for a naturalist scientific foundation of language studies. Schuchardt's main concern was with the ecology of actual language use, with the ways in which a language 'lives' in a community, with the sociological aspects of language and language use, with the actual 'on the ground' mechanisms of language change, with the birth and growth of Creole languages and other topics of that nature. Together with Johannes Schmidt he developed the famous wave model of language change as a sociological process, each individual change of the local vernacular spreading over a well-defined territory. 22 Schuchardt criticised the Young Grammarians as he found their more or less mechanical view of 'exceptionless sound laws' too rigid: close observation of linguistic reality shows that the famous sound laws are not like laws in physics but more like regularities that allow for exceptions according to specific circumstances or preferences among the speech community. 23 Since the social aspect of language was central to Saussure's own thinking, 24 one would have expected some reference to Schuchardt, but there is none. Nor does Saussure seem to have been influenced in any way by Schuchardt's work, no matter how well known and influential it was.
It does not look, therefore, as if Saussure was much concerned with what was going on elsewhere in linguistics or in the wider sphere of philosophical and science-theoretic discussions of his day. On pp. 33-34 of the Cours we read: It is the psychologist's task to determine the exact place of semiology [i.e., the theory of signs; see below; PAMS]; the task of the linguist is to define what makes a language a special system in the totality of semiological facts. The question will be taken up again below. Here we merely emphasise one thing: if, for the first time, we have been able to assign to linguistics a place among the sciences, it is because we have linked up linguistics with semiology. (italics mine) 21 Joseph, Saussure, 254. 22 Johannes Schmidt, Die Verwandtschaftsverhältnisse der indogermanischen Sprachen, (Weimar, 1872). Schmidt does occur in the Cours, briefly, on pages 277 and 287, but only in connection with his wave model of sound change, not in the context of general linguistic theory. Schuchardt is never mentioned. 23 Hugo Schuchardt, Über die Lautgesetze. Gegen die Junggrammatiker (Berlin, 1885). 24 The Cours brims over with references to the social nature of language, a remote reflection not only of Auguste Comte's French positivism, which introduced the notion that societies, much as physical nature, are regulated by laws, but also of the often violent emergence of socialism, communism and anarchism. Two quotes should suffice. We read (Cours, 25): '[The language <ONE speaks>] is at the same time a social product of the language faculty and an ensemble of necessary conventions, adopted by the social body in order to allow for the exercise of this faculty by each individual' and (Cours, 30): '[The language] is a treasure deposited by the practice of speech in the subjects belonging to a community, a grammatical system virtually existing in each brain, or, more precisely, in the brains of a totality of individuals; for a language is not complete in any single brain, it exists perfectly only in the mass of speakers.' The italicised phrase 'for the first time' suggests that Saussure, if he was aware of the attempts by other linguists 'to assign to linguistics a place among the sciences', considered them uselessexcept, of course, Whitney's, whose name is approvingly mentioned a few times, even though Whitney's theorising is hardly of special relevance to Saussure's, much less, for example, than Schuchardt's. 25 It is clear that Saussure felt that the linguistics of his day was insufficiently 'scientific', which made him look elsewhere for the methodological help he needed. This he found, apparently, in the French philosopher-psychologist Victor Egger and the equally French philosopher-journalist Hippolyte Taine, both to be discussed below. Even though Saussure never refers to them by name, their ideas are unmistakably there, all over the Cours. Section 4.1 below supports Aarsleff's contention that Saussure's connection with the more remote hinterland of philosophy and psychology consisted to a large extent in his familiarity with Taine's writings. 26 Taine's aim was to turn psychology into a real science and he did so via his notion of the 'sign', precisely what Saussure aimed at doing with regard to linguistics, witness the quote just given. Egger (see Section 4.2) studied the phenomenon of inner speech and the place of the linguistic sign in that process, clearly steering Saussure towards a purely psychological notion of the sign. Why Saussure never mentioned Taine or Egger is enigmatic, like so much else in Saussure's life. Perhaps he felt embarrassed to admit to his own inadequacy in regard of questions of overarching philosophical scope? Or perhaps he only wished to refer to people he knew personally and had a personal bond with, regardless of their views, such as Whitney or Pictet? We do not know.

The Cours: critical résumé of its main points
I have used the fourth edition of the Cours, from 1949, still published by Payot. According to the editors Bally and Sechehaye, the third edition is, apart from some minor corrections, identical to the second of 1922, which 'does not differ essentially from the first. For the second edition, the editors restricted themselves to matters of detail in order to render the text, at certain points, clearer and more precise' (Cours, 11). The fourth edition carries no special preface; it appeared just after the two editors had died, which allows the inference that no further changes were made. I make no direct use of the otherwise most valuable text-critical editions by Rudolf Engler and Tullio de Mauro, 27 as the additional details supplied there have little or no bearing on the present argument. As for Saussure's private notes and correspondence, and the notes taken by his students, I rely on Godel, 28 de Mauro's 1972 edition of the Cours and Joseph's Saussure.
Rather than follow the Cours chapter by chapter, I restrict this résumé to the points that have provoked most interest and discussion during the decades that followed its publication. For Saussure, linguistics has, in a general sense, three tasks: (i) to describe and trace the history of as many languages as possible, establishing the family relations among them; (ii) to search for the forces that are operative in all languages and define the general laws that all specific phenomena of their histories are subject to; and (iii) to delimit and define itself (Cours,20). Later, on p. 33, it is added that 'the task of the linguist is to define what makes a language a special system in the totality 25 Godel, 'F. de Saussure's Theory of Language' of 1966 opens with a flourishing tribute to Whitney, which, however, contains nothing of substance. All Godel says, as regards the relation between Saussure and Whitney, is: 'Whitney's ideas undoubtedly stimulated de Saussure. He never ceased to feel indebted to the American scholar and many years later, when he offered courses in general linguistics at the University of Geneva, he did not fail to mention Whitney's name with praise and to discuss his ideas.' Given the absence of any specific crucial issue on which both agreed (or disagreed), it looks very much as if Saussure's attitude with regard to the gracious gentleman Whitney was more one of personal loyalty than based on intellectual inspiration. of semiological facts.' Of special relevance is Saussure's insistence (21) that '[a]t bottom, everything in any language is psychological, including its material and mechanical manifestations, such as sound changes.
[ … ] The essential thing about languages, as we shall see, is alien to the phonic character of the linguistic sign.' A specific language is a system consisting of a lexicon and a grammar, largely but not perfectly 'deposited' in the mind of each of its speakers. Any such system is a department of language in general ('le langage'), along with actual speech events ('parole'), which involve physical events (brain processes, muscular movements, sound waves, auditory reception) that take place as a function of the language system.
To explain this, he finds a useful point of departure in what he calls the speech circuit ('le circuit de la parole') (Cours, 27), which consists in the circular process starting with the arising of a 'concept' in the mind of one speaker (A), who converts the 'concept' into sound via the acquired system of his or her specific language ('langue'), followed by the physical transmission of the sound to the ears of the hearer (B), who accesses a corresponding 'concept' and may call up another 'concept', which is then converted into sound, and so on. Figure 1 reproduces Saussure's schematic rendering of this circular process. The purely psychological part consists in the mutual interaction between 'concepts' and 'acoustic images' in the heads (minds) of the participants A and B, as shown in Figure 1b.
One notes that Saussure makes no mention of the fact that there can be no parole, no language use, unless it is about something, either in the real or in a virtual world. 29 What is missing in his analysis of the speech circuit is the fact that, both in the production and in the interpretation of utterances, the participants in the speech circuit are in a mental relation of intentionality with regard to things in a, real or imagined, external world-a relation which must be taken to be mediated by what he calls the 'concept', but this notion is left totally undefined in the Cours or anywhere else in Saussure's notes. This point is taken up in a moment, when Saussure's notion of the linguistic sign is discussed.
It must be kept in mind, meanwhile, that the 'grammatical system' he mentions is, in principle, meant to be a system of word formation or morphology (Cours, 30-31): By contrast, speech ('la parole') is an individual act of the will and of the intellect, in which we can distinguish: 1. the combinations by means of which speaking subjects make use of the code of the language with a view to expressing their personal thoughts; 2. the psycho-physical mechanism allowing them to externalise these combinations.
Saussure, like practically all other authors of his day, places syntax in the 'parole', not in the 'langue', even though he speaks of 'the code of the language'. But this latter notion, which could have been a good starting point for the development of a theory of syntax, remains unmentioned, and thus also undefined, in the remainder of the Cours. Again, on p. 38, the langue is said to 'exist in the collectivity, in the form of a sum of imprints deposited in each brain', whereas parole is 'the sum of what people say, which comprises (a) individual combinations, depending on the will of  Thus, for Saussure, as for the majority of linguists (and the numerous linguistic dilettanti) of the two centuries preceding him, syntax, in the sense of how to combine lexical elements for the expression of any given thought, was not a part of the language system ('langue') but, entirely in the spirit of the Romantic movement, of a free, stylistically creative and aesthetically pleasing combination of concepts and the corresponding lexical elements, that is, of 'parole'. For Saussure, 'grammar' equalled 'morphology' and what we now mean by 'syntax' did not exist.
The notion of syntax as a free, creative, even artistic, combining of words was of a piece with the view, which prevailed during the entire period of romanticism, that human language originated as a form of ritual chant either during collective labour or as part of magical or religious rituals. A minor linguist hesitantly holding this view was Adolphe Pictet, seen by Joseph 30 as the major intellectual influence on Saussure-a thesis that is refuted below. Strictly speaking, Pictet remained noncommittal as to the origin of language, but through his nebulous prose it transpires that he sees ritual chant as lying close to its roots: The first developments of the manual arts, of language, of the social institutions, of the religions, remain enveloped in thick darkness. But it is evident that man cannot in one leap have managed to erect these gigantic edifices whose ruins still strike us with awe, or have clad in a rich and harmonious language the ancient traditions and grave teachings of religion.
[ … ] No doubt, popular chant must have sounded long before the sacred hymn and the heroic poem. 31 Saussure's Cours contains no trace of such a romanticist appeal to chant as lying at the origin of either language or poetry. On the contrary, for Saussure language emerged as a natural product of man's cognitive faculties, entirely in the spirit of Taine. He just had not yet developed an eye for the rule-bound nature of syntactic constructions.
A few examples will show what is meant. Saussure, along with all linguists of his day, was fully aware, for example, of the fact that the vowel change between the English singular foot and its irregular plural feet is an isolated remnant from Old Germanic, where the plural of masculine nouns involved an umlaut, in this case a change from /u/ into /i/ (Cours,110,122), and that the same umlaut is found in the German regular plural Füße for the singular Fuß, though the German umlaut changes /u/ into /ü/. Yet, again along with the entire linguistic world of his day, Saussure had no clear notion of the fact that, for example, the markedly different word order in the English sentence (1) and the corresponding German sentence (2) is likewise a matter of rule-bound constraints within the communities of English and German speakers, respectively, not at all of a free, stylistically creative and aesthetically pleasing stringing together of words: (1) Harold has never wanted to let his children read the Bible.
(2) Harold hat niemals seine Kinder die Bibel lesen lassen wollen. (literally: 'Harold has never his children the Bible read let want') This is clearly a syntactic difference between the two languages, which have rather different complementation systems, that is, systems for the treatment of propositions embedded as arguments under a higher predicate. 32 It wasn't until the 1950s that European linguists began to be aware of the stringency of phenomena of this kind. 33 Since then, developments in the theory of grammar, especially across the Atlantic, have consolidated the notion of syntax as a partly language-specific and partly universally constrained system of rules for the expression of propositional thoughts under a speech-act operator, and thus as part of the grammar of any specific language involved.
As we have seen, a language is, for Saussure, the shared psychological (mental) property of a community, consisting of a lexicon and a grammar (that is, for him, morphology). The lexicon is a collection of signs. Therefore, Saussure consistently concludes, a sign is the shared psychological (mental) property of a community. A sign, moreover, is a psychological element sui generis, in that it links up a concept, also called the signified ('signifié'), 34 with an acoustic image, also known as the signifier ('signifiant'), both being of a psychological nature. Saussure's notion of a linguistic sign is famously visualised as in Figure 2: Linguistic sign systems are but an instance of sign systems in general, such as pantomime, gestural languages, (im)politeness sign systems, military signs, symbolic rituals, etc., but they are at the same time the most important among them (Cours, 33, 101). Sign systems have in common that 'they express ideas' (ibid.)-though no further analysis is provided of what is meant by 'express' or by 'ideas'. The study of sign systems in general Saussure calls semiology (Cours, 33), a field of study yet to be set up but of immense potential importance for a proper understanding of humanity. 35 Linguistics will thus be a branch of the more general discipline of semiology, which falls within the remit of psychology. Since, for Saussure, linguistics is the study of specific languages and of language in general, and since a specific language consists of a lexicon and a grammar, the lexicon being a collection of signs, it follows that the study of signs must be seen as a subdiscipline of psychology. This conclusion is important when we discuss Saussure's relationship with regard to Taine and Egger in Section 4.
After a few centuries of virtual silence, the notion of sign re-emerged after 1870 all over the Western world of philosophy and psychology. In his De l'intelligence, first published in 1870, Taine devotes the first chapter to the notion of sign, soon followed by the French psychologist-epistemologist Victor Egger, who devoted the last chapter of his 1881 PhD thesis La parole intérieure, which dealt with inner speech, to that notion. The somewhat singular American philosopher Charles Sanders Peirce (1839-1914) followed suit: his entire oeuvre is replete with discussions of or about the sign, but Saussure may be excused for not knowing about that, as Peirce was a bit like Saussure The first technical occurrence, to my knowledge, of the term signified (French signifié) as a noun in the Saussurean sense is in Egger's discussion (Victor Egger, La parole intérieure: essai de psychologie descriptive, présentée à la Faculté des Lettres de Paris, [Paris, 1881], 281) on whether a sign precedes the signified (as clouds precede rain), or are simultaneous with (as a red face goes together with shame), or follow the signified (as destruction follows the passing of a cyclone): 'No doubt the sign precedes the signified in certain cases, but in other cases it is simultaneous with it or follows it.' ('Sans doute le signe précède le signifié dans certain cas, mais d'autres fois, il l'accompagne ou le suit.') (italics mine). The term also occurs in Volume I, p. 331, of Hippolyte Taine, De l'intelligence, 2 vols. (Paris, 3rd edition 1878), but only as a past participle adjectivally attached to the noun event (French événement).

35
What Saussure envisaged as the general discipline of semiology never got off the ground, mainly because, during the twentieth century, the notion of sign, though rightly central to Saussure's thought, came under a cloud of disrespect in all branches or schools of linguistics, which wanted no truck with the vaguely esoteric and semi-mystical 'semiotic' movements, such as 'General Semantics', which arose during the past century largely on the basis of Saussure's sensible but unelaborated notion of semiology. Saussure himself never elaborated the general notion of sign any further, though he could have found useful starting points in the Stoic, Augustinian and medieval notions of the sign (see below). In my view, the total neglect of, and the implicit contempt for, the notion of sign during the twentieth century has done great harm to the discipline of linguistics. I have presented a compact statement of my own analysis of the notion of sign in my Language in Cognition, 280-84.
in that he was in the habit of publishing erratically or not at all (his complete writings have been published only recently). Saussure's Cours of 1916 is just one among the many attempts at clarifying the notion. Ogden and Richards' famous 1923 book The Meaning of Meaning comes next, 36 followed by the German psychologist Karl Bühler , with his 'Organonmodell' in his magnum opus Sprachtheorie of 1934, 37 and the American semiotician and philosopher Charles W. Morris , not to mention the doctrines and writings of the theosophical and other occultist movements that flourished at the time-with the result that the very notion of sign came under a cloud of suspicion during the twentieth century (see n. 35). All these authors developed their own notion of what constitutes a sign, though not all of them equally consistently or fruitfully. One cannot say, therefore, that Taine, or Egger, provided Saussure with a notion of sign: their notions differed too much from each other. But one can say that, as from 1870, the notion of sign was at the centre of intellectual interest and was widely discussed all over the Western world.
One reason why linguistic sign systems are considered to be the most central and the most important category of sign systems lies, according to Saussure (Cours, 101), in the fact that the linguistic sign is, in principle, arbitrary-a point widely belaboured in a multitude of works at the time, as, for example, in Egger's La parole intérieure, whose final Chapter 6 is devoted to this notion: In sum, arbitrary signs have a double advantage. First, they are the only possible expression of non-perceptual ideas and of perceptual ideas that cannot be externally represented by an analogous muscular movement.
[ … ] Secondly, they alone can express general ideas as such. 38 There is no intrinsic reason why, for example, what we know as a tree should be called tree in English but arbre in French: the difference is due to mere historical and sociological accident. True, there is quite a bit of sound symbolism or onomatopoeia, such as cuckoo for the bird of that name, in the lexicons of languages, but these languages would function equally well without it. 39 The advantage of the arbitrariness of the linguistic sign, in Saussure's eyes, is that it allows for a much wider range of possible forms than can be realised in systems of motivated signs. All it takes for an arbitrary form to be associated with any given meaning ('concept') is acceptance by the speech community (English has, or used to have, the word acceptation, French acception, for 'commonly accepted meaning'). Arbitrariness thus allows for complex sign systems, such as natural languages, whose rules can be defined regardless of specific forms for the signs used.
The question of whether lexical forms are arbitrary or are, at least originally, naturally motivated goes back to Plato's dialogue Cratylus, where Plato toys with the idea of a natural origin but does not commit himself. Aristotle takes a firm stand: single, non-composite linguistic signs are in principle conventional and arbitrary, not natural or 'motivated'. In Saussure's day, there were still authors, mostly Romantically inspired dilettanti, who defended the natural-origin view, but the professionals were largely agreed that the linguistic sign is in principle arbitrary, as long as its phonological or gestural properties stay within the bounds of possible physical execution.
According to Joseph, 40 the main source for Saussure's notion of the arbitrariness of the linguistic sign was the Geneva-born Adolphe Pictet (1799-1875), a pre-Young-Grammarian comparative linguist and outspokenly romanticist philosopher, like Saussure belonging to the highest Geneva elite and related to the Saussure family. 41 It is true-a fact given great prominence by Joseph-that Pictet stresses the arbitrary character of the linguistic sign, on the rare occasions he speaks of natural language in a general sense: 'In language, sound has, in effect, no longer any immediate sense; it is reduced to the role of arbitrary sign, whose meaning varies infinitely according to the diversity of languages' (Pictet, Du beau, 252). Yet, as rightly observed in Aarsleff, 'Duality the Key', the question of the arbitrariness of linguistic signs is so ancient and has been discussed so widely through the ages (also by Egger, for example, as we have just seen, in Saussure's own age), that one does not need a Pictet to remind one of it. 42 Moreover, for Pictet the arbitrariness of the linguistic sign applies only to the more 'advanced' languages, since, in his view, the words of more 'primitive' languages are somehow 'motivated' in that they reflect an original meaning that got lost through the ages. 43 In other words, Pictet defends the theory proposed by one of the discussants in Plato's Cratylus, according to which names and word forms were originally 'motivated' but their motivation got lost as the forms were transformed through the course of time. In this vein, the Greek word ánthrôpos ('man'), for example, is considered in the Cratylus to be derived from the 'motivated' form 'anathrôn ha ópôpe' or 'one who reflects on what he has seen'. In the same way, Pictet lets a word like English ox be derived from 'Sanscrit ukshan, literally "he who sprays or fertilises" (the cow), from the root uksh, conspergere, effundere (semen)'. 44 Saussure would most certainly not have wished to present his notion of the arbitrariness of the linguistic sign in such terms.
The same goes for the linear character of signs in language, as opposed to those of some other sign systems. The fact that linguistic form is largely composed of linearly ordered elements is due to the nature of the medium used, as the production and perception of sound necessarily involves extension over time. There is, of course, also a simultaneous element in the composition of linguistic sound, in particular in the distribution of accents and intonation, but the main distinctions need temporal sequence for their realisation. This is true, but only trivially so, and nothing much hangs by it. After all, the gestural languages of the deaf make much less use of temporal linearity and much more of visual simultaneity than the sound-based languages of the hearing, a fact that in no way affects their status as natural languages. Pictet also mentions the linearity and articulation of sound, which makes it a particularly useful medium for the expression of the emotional and intellectual 'movements of the soul', as opposed to the materials used in sculpture or painting, which are static and motionless. 45 Yet here again it seems futile to see this as a possible influence on Saussure's thought, for the simple reason that Saussure never touches on this aspect.
Of far greater importance is the fact that Saussure is at pains to stress that the linguistic sign is of an entirely psychological character: it 'unites not a thing and a name but a concept and an acoustic image' (Cours, 98). He explicitly and repeatedly dismisses the notion of a linguistic sign as a term that corresponds to an object or a thing ('une chose'). Above, I have criticised Saussure for his failure to account for the fact that the use of language necessarily involves a relation of intentionality with regard to things in the real or an imagined world, the necessary aboutness of speech. This criticism does not apply directly to his notion of the linguistic sign, though it does indirectly. It does not apply directly since, as we have seen, Saussure maintains that a language ('langue') is a system of lexical signs deposited in the minds of the individuals belonging to a language community. That is, a language exists in the minds of individuals only at type level, the tokens being produced during actual speech ('parole').

42
See also Aarsleff, From Locke to Saussure, 363-64: 'Among today's Saussurian commonplaces are also the linearity of speech and the arbitrariness of the linguistic sign. Taine never expressly discussed either one, but he obviously took both principles for granted [ … ].
[ … ] Saussure's two principles of linearity and arbitrariness did not need to be discovered; they were old and could be found in available texts.' The type/token distinction is already found in Taine: 'A general and abstract idea is a name, just a name, the signifying name for a series of similar facts or a class of similar individuals, normally accompanied by the perceptual but vague representation of some specimen of this class of facts or individuals.' 46 But it was the American philosopher Peirce who introduced the now generally used terms token for unique hic-et-nunc objects or occurrences, and type for the product of the mental process attributing 'sameness' to different tokens reckoned to belong to the same type at any level of abstraction. 47 Tokens are unique and occur in (actual or virtual) reality; types are mental constructs resulting from cognitive identification processes. Mathematically speaking, types are 'characteristic functions', which take concrete, individual tokens as input and deliver a type-membership judgment as value: 'this token does or does not fall under this type.' In this sense, both concepts and acoustic images are types, and thus of a mental nature, and so are the Saussurean signs, which weld the two together.
Saussure, though unaware of Peirce's 'Prolegomena' and his pair of terms, was, like Taine, clearly aware of the distinction between type and token: The problem of identities is found everywhere.
[ … ] Thus we speak of identity with regard to two express trains 'Geneva-Paris, 8.45 pm' leaving at intervals of 24 hours. To our eyes it is the same express train, yet, in all probability, the locomotive, the carriages, the personnel are all different. Or else, if a street is demolished and built up again, we speak of the same street, even though, materially speaking, nothing of the old street may have remained. How can one rebuild a street from the ground up without it stopping to be the same street? Because the entity it constitutes is not purely material. It is based on certain conditions that are alien to its occasional material instantiation, such as its location with regard to other streets. Likewise, what makes the express train the train it is, is the hour of its departure, its itinerary, and in general all the circumstances that distinguish it from the other express trains.
[ … ] The link between two occasions of use of the same word is not based on material identity, nor on exact sameness of meaning, but on elements that will have to be sorted out and will take us very close to the real nature of linguistic units. (Cours, It is clear that Saussure placed the language at type level and speech at token level. And the type level is indeed purely mental, part of our cognitive system that makes us interpret and categorise sensations of token events. Thus far, therefore, the criticism of not taking into account the things we talk about when we talk does not apply. But it begins to apply when we consider that one function of concepts, as type-level functions, is their role in the process of identifying and classifying things we take to be part of the real or of some imagined world. This intentionality or potential aboutness of concepts, and of language in general, is their very raison d'être. A definite noun phrase, such as the girl, is defined in the language system, but it does not refer until actually used in a proper context or situation so that the entity intended by the speaker can be identified by the hearer, if all goes well. Only then do (assertive) sentences acquire a truth value. A sentence like The girl was right after all is a product of the English type-level language system. We can analyse it and identify its subject, its predicate, and the adverbial adjunct after all. But we cannot say whether it is true or false. For that, we need to know what specific girl is meant and what the occasion was at which her opinion or assertion was at issue-that is, the sentence has to be keyed to a specific part of the external world. Concepts alone are not sufficient for this keying process, but they are crucial in assigning conceptual labels to any given entity. Saussure's omission of this intentional aspect of concepts may thus be taken as a point of criticism, but other than that, it is correct to say that, at type level, the linguistic sign is a purely cognitive element uniting a concept with an acoustic image, which is again the type-level representation of token-level sound (or script) events. It is only at token level that we can speak of a linguistic sign being a perceptible form showing the perceiver that a certain concept or a certain proposition has occurred in the speaker's mind. Saussure thus defines the linguistic sign as a type-level mental unit. In this respect, he deviates from the tradition, which has always defined signs as token events functioning as signs on account of their type-level interpretation. I cannot, in this study, trace the long history of the notion of sign through the ages, but I may quote Augustine, whose definition of the sign was mainly inspired by the ancient Stoic philosophers and has been extremely influential through the Middle Ages and after. In his De Doctrina Christiana, at II.1.1, he defines the sign as follows: For a sign is a thing which, besides the form that it presents to the senses, makes something else enter into the perceiver's thought on account of itself. (Signum est enim res praeter speciem quam ingerit sensibus aliud aliquid ex se faciens in cogitationem venire.) In his Dialectica, a few lines down the beginning of Chapter V, one reads: A sign is something which shows not only itself to the senses but also something else to the mind. Much later, Ogden and Richards would refine this traditional notion and present the famous triadic diagram shown in Figure 3, where 'symbol' stands for what is called 'sign' here: 48 Whatever the merits or demerits of these notions of the sign, it is clear that both Augustine and Ogden and Richards define the sign as a token occurrence of a perceptible form, whose interpretation requires type-level mental processing which reveals the existence of 'something else' in the world, Ogden and Richards' 'referent'. But, as has been said, Saussure was free to define the sign at type level, where the relation to elements in the outside world may be left unmentioned, even though it is implicit in the notion of concept.
Not so, however, for the speech circuit shown in Figure 1 above. There it cannot be denied that (real or virtual) world elements and states of affairs are essentially involved: when I ask you to hand me that book, you are addressed as a real person and the idea is that you hand me, also a real person, that particular book, again a real thing. The keying of linguistic constructs to actual or virtual situations and things is the raison d'être of the language system, even though we have not so far discovered what it means to say that mental content is 'intentionally related' to world elements or how exactly this relation of keying is established. One may, therefore, justifiably criticise Saussure for leaving the element of reference to the external world out of account in his analysis of the speech circuit, which involves the occurrence of sign tokens. In Section 4.1 below, this flaw in Saussure's analysis is ascribed to the influence exerted by Taine, who defended a highly idiosyncratic minimalist ontology in which, counter to then current scientific notions, most of the role of physical matter is sacrificed to the complexities of psychological processing-that is, extreme nominalism.
A further distinction for which Saussure has become famous (but see note 1 above) is that between the diachronic and the synchronic study of language. Nowadays, this distinction is obvious to any linguist worth his salt, but one should realise that, at the time, the notion of a system behind the functioning of language was totally alien both to the general public and to linguistic specialists. One was simply unable to see that a language forms a system at any given point in time. The only point of view was historical: only as the product of historical developments could a language be studied scientifically. The leading linguistic authority Hermann Paul  writes in his Prinzipien der Sprachgeschichte ('Principles of the History of Language'): It has been said that there is another scientific way of looking at language than the historical one. I must deny that. What is said to be a non-historical yet scientific way of studying language is in fact nothing but an imperfect historical way [ … ]. The moment one goes beyond the mere statement of isolated facts, the moment one tries to understand the system ['Zusammenhang'] or to understand the phenomena, one treads on the ground of history, even if one may be unaware of that. 49 And again on p. 28: The psychological organisms as described above are the real bearers of the historical development. That which is really spoken has no development at all.
Even so, the Young Grammarian Paul provides much more in the way of syntactic analysis, in the modern sense of the term, than Saussure does in his Cours. He even deals in some detail (Prinzipien, 1920 edition, 282-86) with the great subject-predicate debate that dominated linguistic theorising between, say, 1850 and 1930 but is not mentioned at all by Saussure (see note 16 above). It is, in fact, only by seeing the system that the notion of a synchronic system grows on one. And Saussure himself, as we have seen, still had considerable difficulty actually seeing the system, which, for him, was largely restricted to word morphology. All he could do was talk about it in vague and abstract terms.
For Saussure, the actual acoustic image as fixed in the language system does not matter for the system to function properly, as long as each sign has a more or less unique acoustic image assigned to it. A language is like a game of chess (Cours, 149): all that matters for the game is that each piece is formally different from each other piece and the position of the pieces on the board. As long as one chooses for each piece any old object that fits the board, one can play chess, provided each piece type differs from all other piece types in perceptible form and each piece type has the status defined for it by the rules of the game: it functions as a white or black rook or pawn or knight, etc. Therefore, all that matters for the game is the number of different pieces on either side and a precise definition of what, given the state of the game, each piece can do in the game, its value ('valeur'). A move in the game corresponds to a change in the language.
Saussure thus concludes: 'The linguistic mechanism consists entirely of identities and differences' (Cours, 151). And since identities functioning in systems are taken to be exhaustively definable in terms of differences, Saussure says: All that precedes boils down to saying that in any given language there are only differences. What's more: in general, a difference supposes positive terms between which it can be established. But in a language there are only differences, without any positive terms. Whether you take the signified or the signifier, the language provides neither any ideas nor any sounds pre-existing to the linguistic system, but only the conceptual and phonic differences as they are defined in the system. Whatever there is in the way of ideas or phonic material in a sign matters less than what there is around it in the other signs. The proof of this lies in the fact that the value of a term can be modified without touching either at its sense or at its sounds, but only because some other, neighbouring term has been changed. (Cours, italics original).
Yet the next paragraph presents the reader with the opposite: But to say that all is negative in the language is true only for the signified and the signifier taken separately. As soon as one considers the sign in its entirety, one finds oneself faced with an entity that is positive in its order.
[ … ] Although the signified and the signifier are, each taken separately, purely differential and negative, their combination is a positive fact. It is even the only kind of facts that makes up a language, since linguistic 49 Hermann Paul, Prinzipien der Sprachgeschichte (Halle, 1880) (5th edition 1920); p. 20 of the 1920 edition.
institutions are characterised specifically by the maintenance of the parallelism between these two differential orders. (Cours,(166)(167) The conclusion is, alas, that the language, being purely negative and differential, contains no signs, despite Saussure's repeated assurance that a language is a system of signs.
I shall not try to disentangle this confusion beyond saying that what one sees here is the wellknown phenomenon of a man groping for light and trying to work his way through a dense mass of indistinct and unruly thoughts and ideas which he is as yet unable to control or put in order. Little wonder that Saussure was unwilling to give his courses and even more unwilling to have them published. His confused notion of 'value', though often presented by Saussurologists as the apex of his genius, was never picked up by the world of linguists. No actual, well-elaborated system of 'values' has ever been presented and there is no good reason to try and change that situation. What is more relevant is the fact that we see Saussure here trying again, this time in a different manner, to minimise the material side of language and reduce the whole of language to structures and events that live and take place exclusively in the minds of speakers belonging to a language community. In this respect he was unique, besides being unsuccessful, and in this respect he was in close agreement with the equally idiosyncratic Taine, as is shown in Section 4.1 below.
We now pass on to the central question at issue in the present study: on what sources did Saussure draw as he struggled to formulate a general theory of language? This question has lately given rise to some controversy, which has not, however, created greater clarity. The present study is meant to be a contribution towards that end.

Influences …
Influences indeed. What are they? How can they be shown to exist or not to exist? In most cases they form a more or less indistinct mishmash of education, circumstances, episodes, events, people, books and what not. Yet, leaving the mishmash for what it is, one feels one can often identify major or outstanding intellectual influences, important decisions taken at intellectual crossroads, or thoughts that were picked up, accepted and developed further. Saussure will not have been too different from most in his weighing of authors and ideas. It seems, therefore, that if we want to trace 'influences' in his work, the best we can do is look for major, distinct influences, leaving the indistinct ones to remain buried in the mishmash that lay irretrievably at the bottom of his soul and mind.
This means that we must trace the currents of thought he felt at home in and consciously identified with. And in doing that, we should first of all look for explicit textual evidence of his own decisions and position-takings. Where these fail, we have no choice but to fall back on interpretations based on less direct evidence, consisting either in terminological or textual borrowings or, ultimately, in concordance of thought-thereby leaving ample room for originality on his part. Unfortunately, as we have seen, Saussure was very sparing in mentioning names of people whose ideas he weighed and took on or rejected, which means that we must fall back on the less solid method of plausible conjecture, whether textual or concerning relevant content. What we can do in any case is assign a position to Saussure in the welter of currents and ideas that moved the intellectual world of his day.

The case of Taine
Hippolyte Taine was born in 1828 in Vouziers, in the very north of France, close to the Belgian Ardennes, and he died in Paris in 1893. Although he never had a university appointment, he was a, or the, dominant cultural figure in the French-speaking world and far beyond from 1860 until the end of his life. In 1878 he became a member of the Académie Française. He wrote extensively on history, art, literature and philosophy, but above all on psychology. 'Yet his work does possess its underlying unities-a unity of aim and a unity of method. His lifelong goal was to explore human psychology, [ … ] which he regarded as the very purpose of contemporary thought.' 50 During his lifetime, he was widely known and read. His books sold by the tens of thousands and were continuously being reprinted. His articles were printed in all serious papers and journals and read by everyone with an interest in French culture. 51 But he rapidly slid into oblivion during the twentieth century, to the point of being practically forgotten nowadays. Perhaps one reason for Taine's sudden eclipse from public attention is to be sought in the idiosyncratic nature of his philosophical and psychological theories: though widely read and admired, he never attracted a following-which would have been difficult anyway, as he had no students.
Aarsleff (From Locke to Saussure, 'Duality the Key') conjectures that it was Hippolyte Taine who, though never mentioned, looms large behind Sassure's theorising. The circumstances were right in any case. For Saussure not to have known about Taine, the latter must have been undiscussed in the circles of the Société de Linguistique de Paris, of which Saussure was a very active member, or among the Parisian intellectuals with whom he shared an intensive social life-which is impossible. This, however, is not a sufficient argument. We must also take into account that Saussure never mentions Taine by name, and there is only scant evidence for terminological borrowing. 52 To judge the merit of Aarsleff's conjecture, we must look above all at what I have called the concordance of thought between the two men, in combination with the fact that both were intellectual loners. The real evidence lies in the overall conceptual agreements, with Taine as the leader and Saussure as the follower.
For this, we have to go back to the intellectual climate in France during the mid-nineteenth century. What we find there presents itself as a rich and colourful tangle of thoughts and theories in constant movement and working as much with as against each other but nonetheless showing some dominant strands lending it an identifiable structure.
One such strand was the development of the natural sciences that had started to take wing in the sixteenth century. The natural sciences had developed a highly successful view of the physical world, including the human body, as a system of interconnected 'machines', the well-known 'mechanisation of our world picture'. 53 What happened during the nineteenth century was that the same 'mechanisation' view began to be applied first to human societies, in the work of Auguste Comte (1798-1857) who proposed that societies, like the physical universe, should be seen as complex systems of interacting social mechanisms. Then, in Comte's wake, this view was applied both to history and to the mind in the work of Hippolyte Taine, and, finally, by Saussure and a few others, such as the Polish linguist Jan Baudouin de Courtenay, to language. The mind, in particular, came to be seen as a complex 'machine', to be revealed by hypothesis and experimental testing. 54 This was, in principle, the origin of what came to be called 'structuralism' in some of the human sciences. At least in France, Aarsleff agrees: 'The terminological agreements between Taine and Saussure [ … ] can be extended, though they fall short of the more important conceptual similarities.' (From Locke to Saussure, 362). Among the (few) terminological similarities, Aarsleff mentions (From Locke to Saussure; 'Duality the Key') the simile of the recto and verso of a sheet of paper to illustrate the intrinsic unity of a mental and a cerebral event: '[I]t is clear that they [that is, the mental and cerebral events; PAMS] are like the verso [l'envers] and the recto [l'endroit] of a surface and that the presence or absence of the one infallibly brings along that of the other' (Taine, De l'intelligence, I, 329), and ibid on p. 333: 'The two faces run parallel, and any line that cuts the one also cuts the other at the same height.' Saussure uses the same simile, but with a different pair of terms, for the unity of thought and sound in a language (Cours, 157): 'A language is comparable to a sheet of paper: the thought is the recto ["le recto"] and the sound the verso ["le verso"]; one cannot cut through the recto without at the same time cutting through the verso. Likewise, in a language one cannot isolate the sound from the thought or vice versa. That is possible only by means of abstraction, which would lead either to pure psychology or to pure phonetics.' (One notes that, in the Cours, Saussure otherwise relentlessly stresses the nonmaterial character of what he calls 'sound' here.) One may agree with Joseph, 'Root and Branch', that terminological similarities of this nature are pretty weak evidence. He fails, however, to look at the deeper and more far-reaching currents of thought, as is done in the present study. 53 Eduard J. Dijksterhuis, The Mechanisation of the World Picture (Oxford, 1961). (translation from the Dutch De mechanisering van het wereldbeeld [Amsterdam, 1950]). 54 Others, in particular Taine's contemporary, the German Wilhelm Dilthey , felt that the study of the mind would become scientific not by taking over the 'machine' view but by a deep 'verstehen', an introspectively felt understanding or recognition of what goes on in the mind. This strictly romanticist methodology, which led to an enormous amount of Continental Taine was a key figure in this respect. Émile Durkheim (1858Durkheim ( -1917, for example, the founder of modern sociology, was directly influenced by Taine. 55 In this perspective, Saussure can be seen as the man who tried to do for linguistics what Taine had tried to do for psychology.
Taine elaborated a positivist view of the epistemological problem that runs through modern philosophy: 'The object is not shown to us directly; it is designated to us indirectly by the bundle of sensations that it awakes, or would awake, in us.' 56 His main concern being to lay the foundations for a scientific psychology, he felt that, to achieve this end, it was necessary to reduce mental phenomena to physical events and to detect system in mental phenomena. His De l'intelligence of 1878 is one sustained argument for a token parallelism between mental and cerebral events and processes.
The philosophical question of how to solve the Kantian paradox of knowledge in logico-epistemological terms remains in the background. 57 Nowhere does Taine actually deny the reality of the world, though he comes close. All he is interested in, in this respect, is the fact that reality, whatever it may be, impinges on the senses, creating sensations, and that these sensations are processed by a psycho-cerebral machinery, which claims his full attention. As a result, he tends to take distance from the external world. (Taine is in dialogue with an imaginary opponent): We go even further than you: we think that there are neither minds nor bodies, but simply groups of present or possible movements or thoughts. We believe that there are no substances, but only systems of facts. We regard the idea of substance as a psychological illusion. We consider substance, force, and all the modern metaphysical existences, as the remains of scholastic entities. We think that there exists nothing but facts and laws, that is, events and the relations between them; and we recognise, with you, that all knowledge consists first of all in connecting or adding fact to fact. But when this is done, a new operation begins, the most fertile of all, which consists in reducing these complex into simple facts. A splendid faculty appears, the source of language, the interpreter of nature, the parent of religions and philosophies, the only genuine distinction, which, according to its degree, separates man from the brute, and great from little men. I mean Abstraction, which is the power of isolating the elements of facts and of considering them one by one. (translation by H. van Laun; italics mine) 58 In this passage, whose essence is repeated almost literally in two other works by the same author (see Charlton, Positivist Thought in France, 136), Taine actually says that the least insecure construal of reality is in terms of facts and laws, whereby 'facts' are to be taken as token events of whatever kind and 'laws' as type-level identifications brought about by that 'splendid' cognitive faculty of 'Abstraction'. This ontological agnosticism or scepticism was meant as an answer to the Kantian paradox in that, if the thing in and by itself is unknowable, then we should not pretend to know 'things' and should replace them with unknown values of variables. This agnosticism allows for there to be something 'out there' but remains uncommitted as to the nature or character of that 'something'. We may even allow for 'laws', as Taine did, in addition to 'facts', which leaves room for science to make predictions-the main point of view in the positivism of his day, which remained uncommitted on the metaphysical question of materialism versus 'spiritualism' but reduced all factual knowledge to verifiable perception and generalisation or 'abstraction' (Charlton, Positivist Thought in France, 59-60).
Even though we may feel disinclined to accept Taine's ontological agnosticism, it must be admitted that it was the product of very serious thinking. So serious that it fascinated large sections of the intellectual (and artistic) world, which may not have fully understood the implications but were fascinated. I think that Saussure was one of those who did not fully understand but were European philosophy and, partly doubtful, psychology, including Freudianism, has now been generally discredited as being potentially therapeutic but in principle unscientific. fascinated. And he applied, or, rather, misapplied, this agnosticism to his nascent theory of language. He misapplied it because he failed to realise that speakers of languages have already made up their minds as regards the most viable world construal, which means that language cannot be understood in the abstract terms of 'das Ding an sich' or an unknowable value of a variable but needs, for it to be understood, acceptance of the fact that it functions in terms of reality-as-it-appears-to-us.
With this in mind, it does not seem too far-fetched to hear the echoes of Taine's metaphysical leanings in Saussure's insistence that any given language ('la langue') is of a purely psychological nature, that linguistic signs have no relation to the external world but are merely cognitive units consisting of two equally cognitive elements, a concept and an acoustic image, that a language system functions in terms of variables, not of (images of) reality, that even in actual speech the world objects referred to and the properties mentally assigned to them in the propositions expressed play no rolethe more so because Taine's attempt at metaphysics did not represent a mainstream philosophical trend of thought but remained more or less a private theory, no matter how widely his texts were read and discussed. Taine and Saussure are thus seen to be connected by their common desire to minimise the material side of the universe and maximise the part played by human cognitionthough Saussure was by far the lesser of the two in power and depth of thought.
The application of the 'machine' view to the mind had been expressed earlier, notably by Julien Offray de la Mettrie (1709-51) and Franz Joseph Gall (1758-1828, but Taine differed from these predecessors in one important respect. La Mettrie and Gall held that mental and physical processes stand in a one-to-one relation to each other, forming double-faced but identical systems-which enabled them to deny the existence of the 'soul' or 'mind' as a separate entity (and got both of them into serious trouble with religious and state authorities). Taine, by contrast, was beginning to think, in a tentative and not fully explicit way, of what is now known as token physicalism, the idea that mental processes have an autonomous systematicity, not reflected in the physical, cerebral events to which, in isolation, they stand in a direct and immediate relationship. In other words, we see in Taine's texts the first, tentative, undoing of the ruthless reductionism found in La Mettrie and Gall: while cognitive token events may be taken to correspond one-to-one to token events in the brain, cognitive systems are not reflected as such in the brain. If that were so, psychology would not be a field of enquiry of its own. It would, like La Mettrie's soul, vanish into thin air, or rather into physics. On this essential point, Taine was ahead of Comte, who still endorsed Gall's phrenology and rejected psychology as an autonomous science (Charlton,Positivist Thought in France,36,.
As Jerry Fodor put it on p. 2 of his book The Language of Thought, under strict reductionism: Insofar as psychological explanations are allowed a theoretical vocabulary, it is the vocabulary of some different science (neurology or physiology). Insofar as there are laws about the ways in which behavior is contingent upon internal processes, it is the neurologist or the physiologist who will, in the long run, get to state them. (italics original) 59 Fodor draws an illuminating parallel with systems of monetary exchange (Language of Thought, 15-16). Physically speaking, money cannot be defined in a way that makes sense in physics. Money consists materially in coins, paper bills, bank accounts as they are kept in banks (nowadays as computerised files) and so on. Taken together, these forms of monetary existence do not constitute a category that makes sense in physics. Though token events of monetary exchange will always correspond to some physical token event or series of events, physical science will never be able to provide systematic descriptions and analyses of the nature, function or 'value', of such monetary exchange events, or of the laws or tendencies that are operative in economic systems. For that, we need an autonomous science, called economics. In the same way, we need an autonomous psychology to analyse and explain psychological token events, and an autonomous linguistics to do the same for linguistic token events. Nowadays, the autonomy of cognitive systems is fully recognised, as is their bit-by-bit dependence on, or correspondence with, physiological states and token events in the brain, but the question of how cognitive systems, including the system or systems underlying the use of language, emerge from, or are 'supervenient' on, physiological brain structures and processes, while at the same time influencing and directing them, is still far removed from a satisfactory answer. The autonomy of cognitive systems is, in fact, the central theoretical problem of modern cognitive science.
Next to the towering figure of the German psychologist Wilhelm Wundt , the father of modern experimental psychology, Taine is a key figure in the transition from the fully reductionist stance of men like La Mettrie and Gall to the modern structuralist notion of autonomous cognitive systems. Let us have a closer look at what Taine actually says about this. The inseparable unity of mental and physical token events is stressed, for example, on p. 329 of vol. I of De l'intelligence: This is sufficient reason to admit that cerebral and corresponding mental events are at bottom single events with two faces, the one mental, the other physical, the one accessible to consciousness, the other to the senses.
Yet two pages later, on p. 331, we see how Taine, still in a tentative way, ascends from token events to systems, which no longer stand in a one-to-one relationship with physiological systems: In this distinction [between mental and physical event; PAMS] and in this union, all the advantage is for the mental event. It alone exists. The physical event is nothing but the way in which it affects, or could affect, our senses. For the senses and the imagination, the sensation, the perception, in short the thought, is nothing but a vibration of cerebral cells, a dance of molecules. But that holds only for the senses and the imagination. In itself, the thought is a quite different thing, defined only by its own elements. If it takes on a physiological appearance, that is only because we have translated it into a foreign language, where it takes on an appearance that is not its own.
In fact, whenever Taine speaks of the parallelism of mental systems with physiological facts, it is always in terms of a simile, never in terms of one-to-one correspondence. Thus we read on pp. 175-76 of vol. I of De l'intelligence: At the present day, psychology is faced with what look like simple sensations, just as chemistry, in the beginning, found itself faced with what looked like simple substances. In fact, observation, whether through the senses or through self-reflection, at first grasps only composites; its task is to decompose these into their elements, to show which compositions the elements allow for so that it becomes clear how the diverse composites have come about.
Then, on p. 188: There thus comes about in us an infinite underground work, of which only the products come to our conscience, and that only in rough outline. As regards the elements and the elements of the elements, these are out of reach for consciousness. We come to them by reason. They are to sensations what the secondary molecules and the primitive atoms are to bodies. We only have an abstract idea of them and what represents them for us is not an image but a notation.
The same idea is applied to history and to the structure of societies: There is then a system in human sentiments and ideas; and this system has for its motive power certain general traits, certain marks of the intellect and the heart common to men of one race, age or country. As in mineralogy the crystals, however diverse, spring from certain simple physical forms, so in history, civilisations, however diverse, are derived from certain simple spiritual forms. The one[s] are explained by a primitive geometrical element, as the others are by a primitive psychological element. (translation by H. van Laun; italics mine) 60 And on p. 19 of the same volume: 60 Taine, English Literature, I, 7-8.
Just as in its elements astronomy is a mechanical and physiology a chemical problem, so history in its elements is a psychological problem. There is a particular inner system of impressions and operations which makes an artist, a believer, a musician, a painter, a wanderer, a man of society. (transl. by H. van Laun; italics mine) In the Introduction to his Essais we read: '[M]an is not an assemblage of contiguous pieces, but a machine with ordered wheels; he is a system, not a heap.' 61 And on p. x: '[psychological faculties] form a system like an organised body'. Aarsleff (From Locke to Saussure, p. 361) quotes from the same Introduction: 'History is basically a problem of psychological mechanics', concluding that '[i]n this sense Taine is a true structuralist.' Saussure never touches on the deeper and wider philosophical and theoretical aspects of the mind-body problem, but the echoes of Taine's 'psychological machine' view of the human mind in his Cours are manifold, even though, as has been shown, his notion of the machinery of grammar did not include what today is considered its most central part, namely syntax. Saussure continually stresses the fact that a language is a system of signs or a mechanism: Cours, 30, 33, 40-43, 101, 107 ('this system is a complex mechanism'), 116 ('a language is a system of pure values'), 124 ('a language is a mechanism that continues to function despite the deteriorations it has to suffer'), 140, 149, 151 ('the linguistic mechanism'), 157, 159-63, 176-85.
While Taine regarded the whole of the human mind as an autonomous psycho-physiological mechanism, Saussure did the same with regard to any given language as part of the human mind. Since they were both unique, in their day, in holding these views, and since both were simultaneously active in Paris, where Saussure cannot have failed to know about the ideas of the famous Hippolyte Taine who was no less than a celebrity, the conclusion that Saussure was inspired and influenced by Taine as regards the 'mechanisation' view of the mind seems inescapable. 62 The difference between their common 'mechanism' view on the one hand and their common ontological banning order on physical matter on the other is that the former engendered the short-lived European structuralist tradition in linguistics, carried by Jan Baudouin de Courtenay , Saussure, Antoine Meillet (1866-1936), Nikolai Trubetzkoy (1890-1938), Louis Hjelmslev (1899-1965 and many others, whereas the latter never left a trace in the academic agenda of the twentieth century.

The case of Egger
Victor Egger , born and deceased in Paris, eldest son of the professor of Greek at the University of Paris Émile Egger, received his PhD there in 1881 and taught at the universities of Bordeaux and Nancy from 1877 to 1904, when he was appointed full professor of philosophy and psychology at the University of Paris. Unlike Taine, he was not a public figure but a quiet professor avoiding the public limelight. Joseph (Saussure, shows that Saussure was a highly interested and critical reader of Egger's work. Egger's 1881 book on internal speech shows an author who has fully understood the Cartesian-Kantian paradox of knowledge, discussing it at length in his first few chapters (speaking of the 'nostrum' versus the 'alienum' on p. 104 of the book). Against that background he describes and analyses the phenomenon of inner speech, setting it off in great detail from 'outer' speech and assigning it a place of importance in the speech process. His method is descriptive, based on introspective reports of his own experiences. He stresses the feedback self-screening function of inner speech before an utterance is actually realised-an innovative perspective not found elsewhere in the literature of his day (and not taken over by Saussure), but an important topic in present-day psycholinguistics. 63 Distinguishing between a first, provisional, and a final formulation for a given thought in inner speech, he writes: If, by a stroke of luck, there are no differences [between a first and a last inner speech formulation; PAMS], if the two thoughts coincide perfectly, the provisional expression is accepted as definitive and the new thought has found its true formula immediately. This is what happens when one has inspiration. [ … ] If, by contrast, the two thoughts are found to be different, the provisional formulation is, by that very fact, condemned.
[ … ] The thought that interests us, our real thought, had thus not found the expression it deserved. 64 The last chapter of the book (241-321) is devoted to an analysis of the notion of linguistic sign, with abundant attention for its arbitrariness, as mentioned. One gathers that a sign in general is taken to be anything experienced that shows the experiencer the existence of something else not also experienced. This enables him to say (245) that, for the philosopher but not for the ordinary layperson, 'Thoughts are the signs of thought.' For the ordinary lay person, thoughts are given as such, without the extra 'meaning' that, therefore, humans possess a function or faculty called 'thought': 'One must, therefore, distinguish between two kinds of image, on the one hand those images that are signs and on the other those which are there on their own account and which alone are, properly speaking, thoughts.' A linguistic sign is (248), 'a word attached to an idea', whereby the 'word' may be a type or a token, and if the latter, it may occur in inner or outer (external) speech. Egger thus accepts the notion of 'inner signs' (274).
When the sign has to function socially, it must be external, or material (278): 'A social sign is a common sign, and a common sign is a material sign.' But the 'material' sign is merely accidental, of marginal importance. The sign's main function, one gathers, is internal, in inner speech: Inner speech seems to us to be in us merely the most external clothing of that most subtle and hidden thing we call 'thought'. [ … ] For example, if the sound horse, in inner or outer speech, is the most apparent sign of the idea awoken by it, behind it hides another sign of the same idea, more complex and more direct, which, upon analysis, decomposes itself into the following elements: the visible form of the animal, the clatter of its hoofs on the ground, its whinnying, the image of its harness, that of a rider or of a carriage. 65 Clearly, this notion of linguistic sign is not identical to Saussure's, but one notes the overall drive to eliminate or marginalise all physical elements and focus on the psychological aspects. Since it is known from Saussure's private notes (Joseph,Saussure, that Saussure was an avid and critical reader of Egger's book on inner speech, the inference that he formed his own ideas regarding the linguistic sign in his intellectual struggle with Egger's text seems fully justified.

The case of Pictet
Let us now look more closely at John Joseph's conjecture, in his two largely overlapping publications 'Pictet's Du beau' of 2003 and 'Root and Branch' of 2004, that Saussure's fellow Genevan and relative Adolphe Pictet, mentioned earlier, was a major source of inspiration for him. Pictet was one of the many followers of the romanticist tradition. In Chapter 15 of his Du beau, entitled 'Le classique et le romantique' (279-304), he explicitly takes the side of the romanticists, against the 'classics', his term for what are usually called rationalists and are represented nowadays, in psychology and linguistics, by cognitive scientists and formally oriented linguists. On p. 279 of that book he even speaks of an open row ('querelle') between the two schools or currents of thought. He accuses contemporary psychology of maintaining that the experience of the beautiful consists in nothing but the stimulation of certain nerve centres in the brain. Instead, he wants to analyse the experience of beauty by treating it as composed of more basic experiential units. In more modern terms, he could be called a phenomenologist. 66 64 Egger, Parole intérieure, 225-26. 65 Ibid., 242. 66 He was also an undisguised racist, unlike the majority of romanticists of his age, who fought successfully for the abolition of slavery and for universal human rights. On p. 117 of Du Beau, Pictet draws an ascending scale from the ape to the negro to the European, saying that 'the negro tends more towards the animal than the European type does'. His Les origines indo-européennes starts off with a lofty song of praise for the Indo-European race, 'destined by Providence to rule one day the entire globe' and is 'privileged among all other races by the beauty of its blood, the gift of its intelligence', etc., etc.
In the case of Pictet, there is explicit negative evidence for any influence on Saussure, coming straight from the horse's mouth. Saussure refers to Pictet twice in the Cours, first on p. 297, where he is unambiguously criticised and where the central element in Pictet's theorising is explicitly rejected on the grounds that it has 'obscured' a proper insight into the 'primitive vocalism', that is, the system Saussure himself worked out in his Mémoire of 1879: In his Origines indo-européennes [ … ] A. Pictet, while explicitly recognising the existence of a primitive people that spoke its own language, is nonetheless convinced that one should before all else look at Sanskrit, whose testimony has greater value than that of several other indo-european languages taken together. It is this illusion that has, for many years on end, obscured questions of prime importance, such as that of the primitive vocalism. (Saussure,Cours,297) This passage is not mentioned by Joseph. The second reference is on pp. 306-7 of the Cours, where, no doubt because Pictet's name was still echoing locally in Geneva, Saussure devotes a small part of his lectures to 'linguistic paleontology', Pictet's innovation. Here he not only speaks about Pictet with reserve, but in fact explicitly rejects the man's views: Adolphe Pictet [ … ] is above all known for his book Les origines indo-européennes (1859-63). This work has served as a model for many others; it is still the most engaging of all. Pictet wants to find in the testimonies provided by the Indo-European languages the basic features of the civilisation of the 'Aryâs', and he believes that he can determine their most diverse aspects: material objects (tools, weapons, domestic animals), social life (was it a nomad or an agricultural people?), family, government; he tries to determine the cradle of the Aryâs, which he places in Bactria; he studies the flora and fauna of the lands they inhabited. It is the most elaborate attempt in this direction that has so far been made. The science thus inaugurated received the name of linguistic paleontology. Other attempts have since been made in the same direction.
[ … ] Well now, it seems that one cannot expect information of this kind to be derivable from a language. And if that is so, it is due to the following causes: [ … ] Whereupon Saussure mentions (Cours, 308) (i) the uncertainty regarding etymologies-an implicit criticism of pre-Young-Grammarian comparative philology, already severely condemned on p. 17 of the Cours-giving as an example the erroneous connection made by earlier comparativists between Latin servus ('slave') and servare ('guard'), which led to the unwarranted conclusion that the original function of slaves was to guard the house; (ii) the constant changes of word meanings, combined with the fact that words in related languages or dialects may disappear or be replaced by other words, which does not mean that the people that lack the word or use the new word no longer know of the thing signified by the old word; and (iii) the fact that, for whatever reasons, words are borrowed from one language into another. It is thus quite clear that Saussure was highly critical of Pictet's pre-Young-Grammarian views and attitudes.
According to Joseph, however: The book by Pictet that so marked the young Saussure, Les origines indo-européennes ou les Aryas primitifs, published in two volumes in 1859 and 1863, achieved widespread renown and is the work for which Pictet is best remembered. While Saussure's memoir makes clear that, looking back at the age of 46, he sees his early fascination with the book as juvenile, it is just as clear that Pictet's work remains present and vivid in his mind. Near the end of Cours he says of it that it 'has served as a model for many others; it is still the most engaging of all.
[ … ] His is the most important undertaking of its type.' 67 And: The have here a barely twenty-year-old but intellectually precocious youngster, writing in a local journal of his extremely class-conscious home town Geneva about the second, posthumous, edition of a book by a Genevan upper-class celebrity who had deceased three years earlier, fifty-eight years his senior, closely connected with his own upper-class family, a man he had greatly admired as an adolescent, until his own university studies had made him see the elder man's inadequacies. Could this young man have done anything else but write in laudatory terms about the revered old man? Doing otherwise would have made him an outcast in his hometown. There is no link between language and aesthetics in any of Saussure's work or any of his notes. There is one, a clear one, in Pictet's works, and this link has nothing to do with any 'almost conscious dream of an ideal humanity', as fantasised by the young Saussure in an easily forgivable attempt to save Pictet's face, and everything with the romanticist idea that language originated as poetic chant, an idea which, as explained above, was widespread during Pictet's day but considered of no consequence by the vast majority of professional linguists, including the Young Grammarians and Saussure himself. 72 The remainder of Joseph's 'Pictet's Du beau' and 'Root and Branch' consists of gratuitous exegesis. As has been observed (n. 69), he likens Pictet's rapturous remarks on the beauty of stalactites in damp caves to Saussure's analysis of the structure of language. He likewise conjures up a parallel between, on the one hand, the animal's emerging free will, as exaltingly described in Pictet's Du beau, p. 40-opposed to the lack of free will in crystals and plants-and Saussure's parole ('speech') on the other, as opposed to the langue: [ … ] parole, the individual's wilful speech production, is certainly comparable to the animal. What about langue? It seems to have features of both the crystal and the plant; though in a number of respects the distinction Pictet draws between crystal and plant resembles the one Saussure will draw between langue (a particular language system) and faculté de langage (the overall human language faculty).
[ … ] Besides these parallels in content, the very way in which Pictet imagines and describes the crystal cave and other natural realms prefigures the whole 'aesthetic' of the structural system of langue as Saussure will describe it. 73 On p. 376 of his 'Pictet's Du beau', Joseph uses Pictet's description of how ordinary, as opposed to figurative or poetic, language works as a parallel to Saussure's description of the 'signified' as a concept. For Pictet, the objects that are brought before our eyes in poetry need not correspond to actual reality but pass through a process of imagination. Not so for ordinary language: This is where figurative differs from ordinary language. In the latter, the words address themselves to the intelligence without any intermediary, the very moment they designate real things, because the thought goes straight to its goal, neglecting as a mere accessory the perceived appearance of the object. 74 For Joseph, this means that 'the normal condition of language is one in which the signified is pure thought, divorced from any visual image'. No matter whether this is true, false or uninterpretable, the relation with Saussure is obscure, to say the least.
All this makes it impossible to maintain that Pictet was a major influence on Saussure's theoretical thinking. Joseph states ('Pictet's Du beau', 377) that 'Pictet's Du beau is a key 'missing link'. As such, its status is supported by direct documentary evidence, starting by what Saussure himself wrote about it, and including his own belief that it was directly connected to Pictet's 'linguistic paleontology', but in reality the documentary evidence, including Saussure's own notes, shows the opposite. Aarsleff's rebuttal of Joseph in his 'Duality the Key' is entirely justified.
A final word on Joseph's denial of any influence by Taine. Joseph writes:
vocabulary. Saussure's library did, however, contain a copy of Antinomies linguistiques, with a handwritten dedication to him from Henry. 77 Finally, on p. 508: Nowhere in Saussure's lectures, publications or unpublished papers are the names of Durkheim or Tarde [Gabriel Tarde, 1843-1904; PAMS] to be found. Consequently, as with so many other figures he never cites, their purported influences on him are inferences based on shared terms such as 'collective consciousness' that could well have come via an intermediary source [ … ].
So what is given to others is denied to Taine, even though Saussure is much more likely to have been familiar with Taine's work than with, say, the works of Stallo or Pareto. Given the evidence, together with the intellectual climate of the day and age, it is hard to deny that Taine looms large behind Saussure, as he did behind Durkheim and quite a few others in related disciplines.

Epilogue on methods of argumentation
This essay had its beginnings in the long-standing and bitter dispute between Aarsleff and Joseph about Saussure's intellectual background. In the absence of hard evidence concerning the genesis of Saussure's late ideas, the best that can be hoped for is to establish their context with such accuracy that we may at least highlight the obvious and exclude the fanciful. Following in Hans Aarsleff's footsteps, I have tried to exclude some of the most fanciful suggestions and to highlight Taine as probably the most significant presence in Saussure's linguistic thought. In this mode of argumentation, even more than in disputes based on solid facts, a certain amount of reasonableness is required, and while Aarsleff himself may not always have been as patient with the objects of his critique as one might have preferred, he has in fact repeatedly been denied some of the most elementary conditions of reasonable debate. I feel that a strong case can be made for the thesis that Taine was one of the fathers of twentieth-century structuralism in the human sciences and that he was a, or the, major influence on Saussure's mature linguistic thinking, even if that diminishes somewhat the role and importance of Saussure, the perennial doubter who could not get his ideas into a publishable form and had the bad habit of not mentioning his sources. This would not only sharpen our understanding of the period and of Saussure as a figure in the history of linguistics, but also, not unimportantly, give due recognition to Hans Aarsleff as a great historian of philosophy and of linguistics.