The short answer is no, Saussure did not say meaning is arbitrary.
Why do we care what Saussure said? Because some influential work in cultural sociology makes the consequential (and I think incorrect) claim that meaning is arbitrary and uses Saussure’s work to justify these claims. Consider, as an example, some of the work of Jeffrey Alexander. When the “strong program” of cultural sociology was just a twinkle in Alexander’s eye, he wrote (1990:536):
Since Saussure set forth semiotic philosophy in his general theory of linguistics, its key stipulation has been the arbitrary relation of sign and referent: there can be found no “rational reason,” no force or correspondence in the outside world, for the particular sign that the actor has chosen to represent his or her world.
A few years later, in the strong program’s foundational article, Alexander and Smith claim (1993:157):
Because meaning is produced by the internal play of signifiers, the formal autonomy of culture from social structural determination is assured. To paraphrase Saussure in a sociological way, the arbitrary status of a sign means that its meaning is derived not from its social referent—the signified—but from its relation to other symbols, or signifiers within a discursive code. It is only difference that defines meaning, not an ontological or verifiable linkage to extra-symbolic reality.
Then finally, a more recent example, Alexander writes in Performance and Power (2011:10, 99):
A sign’s meaning is arbitrary, Saussure demonstrated, in that “it actually has no natural connection with the signified” (1985:38), that is, the object it is understood to represent. Its meaning is arbitrary in relation to its referent in the real world…
Not long after Durkheim’s declaration, and quite likely in response to it, there emerged a dramatic transformation in linguistic understanding that continues to ramify in the humanities and the social sciences. Ferdinand de Saussure and Roman Jakobson propose that words gain meaning not by referring to things “out there” in the real world, but from their structured relation to other words inside of language.
In my forthcoming paper, “Becoming A Dominant Misinterpreted Source,” I show that much of this received understanding of Saussure misses the mark.
To begin my journey down the Saussurean rabbit hole, I reviewed 167 articles and book chapters in sociology that cite Saussure to distill the most common interpretations of his work. The figure below shows the pages (on the x axis) of The Course in General Linguistics (Cours) and the number of citations to that page number as a count (on the y axis). Of the 167 citations, however, only 35 offer page numbers. Furthermore, of those offering page numbers, they are mostly confined to four basic topics: (1) the langage, langue, parole distinction, (2) the definition of “semiology,” (3) the definition of the “linguistic sign,” and (4) the definition of “linguistic value.”
What is not cited is over half of the book: Saussure’s discussion of grammar, principles of articulation, diachronic (i.e. evolutionary) linguistics, geographic linguistics, and retrospective (or historical/ anthropological) linguistics. (And, of course it covered this wide range of topics because it was lecture notes for his linguistics course compiled and published after his death.)
Next, to determine if these common interpretations are correct, I engaged in an exegesis of the Cours, as well as reading other text written by Saussure, and also text about Saussure written by his biographers and other linguistic historians. While there are some things we’ve been getting right, there are important things we’ve been getting wrong.
First, it is commonly assumed by sociologists that Saussure was putting forth a philosophy of language — or how language refers to things in the world (often encompassed as the “problem of reference”). He was, however, putting forth a philosophy of linguistics, or how language was to be studied as a science (and, in fact, spends very little time discussing “semiology,” which he saw as a branch of general psychology). The implication of this is that Saussure’s “key stipulation,” as Alexander asserts, was not “the arbitrary relation of sign and referent.” Rather, for Saussure the linguistic sign was a wholly psychological entity, rendering both the physical sound and the physical referent outside the scope of general linguistics.
Saussure claimed that a linguistic sign was composed of two aspects. The first was the mental impression of the sounds of speech (image-acoustique or sound-image), which he called the signifier. The second was an idea or concept, understood in psychological terms, which he called the signified. What was arbitrary for Saussure was not the relation between a spoken word and its referent; rather, what he claimed was arbitrary was the relation between signifier and signified, both mental entities (see Table 1). This arbitrariness, he asserted, allowed the linguist to justify studying the totality of these sound-images as if an autonomous system.
|Mental Entity||Sound-Image (Signifier)||Concept (Signified)|
|Physical Entity||Sounds of Speech||Referent|
Here, we can see a kind of ur-argument for claiming that some object of study is autonomous, and thus requiring the specialized tools of a distinct enterprise. This, I would argue, is why Alexander wants to borrow Saussure for non-linguistic domains. It offers a means to assert that “the formal autonomy of culture from social structural determination is assured.” However, Saussure was very clear that he saw language as a unique entity, and thus his argument for autonomy was also unique to language. Although he acknowledged some ways language was not arbitrary, and sketched out how the study of language was a subfield of the general science of “semiology,” he felt language was set apart by being the most arbitrary of all ( 2009:88):
In order to emphasise that a language is nothing other than a social institution, Whitney [a famous American linguist] quite rightly insisted upon the arbitrary character of linguistic signs. In so doing, he pointed linguistics in the right direction. But he did not go far enough. For he failed to see that this arbitrary character fundamentally distinguished language from all other institutions.
A final, and kind of tricky, misinterpretation of Saussure relates to his definition of “value.” It is often assumed that what Saussure meant by value was synonymous with “meaning.” But “linguistic value” was about the organization of sound-images in the mind, and distinct from the organization of meaning which had to do with concepts or ideas. Furthermore, value is not the same as the qualities of physical sounds, but rather was about how sound-images were related to each other. As Saussure states,
Proof of this [that value is distinct form meaning and physical sound] is that the value of a term may be modified without either its meaning or its sound being affected, solely because a neighboring term has been modified” (Saussure,  2009, p. 120)
Here, we see a second ur-argument emerge, related to Endogeneity and Mutual Constitution. The object of inquiry is not only autonomous, but the components of some system can only be understood with how they relate to every other component in that system. Change one element in the system, and every element in the system changes accordingly. Here again Saussure is quick to argue that language — specifically understood as the system of linguistics values — is unique (Saussure,  2009, p. 80):
…language is a system of pure values which are determined by nothing except the momentary arrangement of its terms. A value—so long as it is somehow rooted in things and in their natural relations, as happens with economics (the value of a plot of ground, for instance, is related to its productivity)—can to some extent be traced in time if we remember that it depends at each moment upon a system of coexisting values. Its link with things gives it, perforce, a natural basis, and the judgments that we base on such values are therefore never completely arbitrary; their variability is limited. But we have just seen that natural data have no place in linguistics.
The final misunderstanding involves whether Saussure is developing Durkheim’s thoughts about culture. To use Alexander again, consider (1988:4–5):
Saussure depended… on a number of key concepts that were identical with the controversial and widely discussed terms of the Durkheim school. Most linguistic historians (eg. Doroszewski 1933:89-90; Ardener 1971:xxxii-xiv), indeed, have interpreted these resemblances as evidence of Durkheim’s very significant influence on Saussure… The echoes in Saussurean linguistics of Durkheim’s symbolic theory are deep and substantial. Just as Durkheim insisted that religious symbols could not be reduced to their interactional base, Saussure emphasized the autonomy of linguistics signs vis-a-vis their social and physical referents.
In the paper, I go into detail demonstrating why this is very unlikely, but here I’ll just quote a couple linguistic historians. The first essay on the matter in English states (Washabaugh 1974:28):
Most linguistic historians (Doroszewski 1933; Ardener 1971; Robins 1967; Mounin 1968) have interpreted these resemblances as evidence of Durkheim’s influence over Saussure. However, a careful reading of Durkheim will show that these resemblances are only terminological.
Perhaps the most comprehensive discussion of the Saussure-Durkheim link comes from Koerner, where he concludes (1987:22): “I do not see… any convincing concrete, textual, evidence that Saussure incorporated Durkheimian sociological concepts in his theoretical argument.”
Is Meaning Really Arbitrary?
A far more important question than whether Saussure actually claimed meaning is arbitrary is whether meaning actually is arbitrary.
Alongside appeals to the authority of Saussure are “just so” stories that seem to show arbitrariness as an obvious fact. As it relates to the present-day Latin alphabet in English, for example, we can assert that the letter “A” is arbitrarily related to the sound that it might represent. However, what about the “O” which does correspond to the shape of the lips when we make the /o/ sound? For the same reason I cannot use this latter example to assert that all letters in the alphabet correspond to the shape of the mouth, we should not use the former to claim that all letters are arbitrarily related to their sounds. Even worse is using such examples to make claims about the operation of meaning in general (the fallacy of composition). The range of arbitrariness or motivation in semiotic systems is, after all, an empirical question which scores of scholars have been exploring for decades. More problematic than misinterpreting Saussure, then, is wielding his lecture notes as a means to shut down this line of inquiry.
Often in tandem with claims that meaning is arbitrary is the assertion that meaning is “conventional,” as if the latter is a prerequisite for, or proof of, the former. But, does this need to be the case? I would argue it does not, and furthermore that this opens up a much broader scope for cultural analysis. The meaning of say, smoke, can be “motivated” in that it is correlated with the presence of fire—but, and this is key, fire is not the only thing with which smoke is associated. As fire is also used to cook, for example, smoke is also associated with food. How do we know whether smoke “means” fire or food if not through some human selection and convention? That the associations between meanings and signs are made more or less probable by the structure of reality does not mean they are not also conventional. Furthermore, I would contend, a more fruitful point of departure for cultural analysis is a framework which can account for both the arbitrary and motivated aspects of meaning.
Alexander, Jeffrey. 2011. Performance and Power. Polity.
Alexander, Jeffrey C. 1988. “Culture and Political crisis:‘Watergate’and Durkheimian Sociology.” Durkheimian Sociology: Cultural Studies 187–224.
Alexander, Jeffrey C. 1990. “Beyond the Epistemological Dilemma: General Theory in a Postpositivist Mode.” Pp. 531–44 in Sociological Forum. Vol. 5. Springer.
Alexander, Jeffrey C. and Philip Smith. 1993. “The Discourse of American Civil Society: A New Proposal for Cultural Studies.” Theory and Society 22(2):151–207.
Koerner, E. F. Konrad. 1987. On the Problem of“ Influence” in Linguistic Historiography. John Benjamin.
Saussure, Ferdinand de.  2009. Course in General Linguistics. edited by A. S. C. Bally. Chicago: Bloomsbury Academic.
Stoltz, Dustin S. Forthcoming. “Becoming A Dominant Misinterpreted Source: The Case of Ferdinand De Saussure in Cultural Sociology.” Journal of Classical Sociology.
Washabaugh, William. 1974. “Saussure, Durkheim, and Sociolinguistic Theory.” Archivum Linguisticum 5:25–34.