While speech (parole) is heterogeneous, that is to say composed of unrelated or differing parts or elements, language (langue) is Homogeneous, composed of the union of meanings and 'sound images' in which both parts are psychological. Therefore, as langue is systematic, it is this that Saussure focuses on since it allows an investigative methodology that is rooted, supposedly, in pure science. Beginning with the Greek word ‘semîon’ meaning 'sign’, Saussure names this science semiology: ‘a science that studies the life of signs within society’.
The focus of Saussure’s investigation is the linguistic unit or sign.
The sign (signe) is described as a "double entity", made up of the signifier, or sound image, (signifiant), and the signified, or concept (signifié). The sound image is a psychological, not a material concept, belonging to the system. Both components of the linguistic sign are inseparable. The easiest way to appreciate this is to think of them as being like either side of a piece of paper - one side simply cannot exist without the other.
The relationship between signifier and signified is, however, not quite that simple. Saussure is adamant that language cannot be considered a collection of names for a collection of objects (as where Adam is said to have named the animals). According to Saussure, language is not a nomenclature. Indeed, the basic insight of Saussure's thought is that denotation, the reference to objects in some universe of discourse, is mediated by system-internal relations of difference.
In Figure 2 above, the signified "tree" is impossible to represent because the signified is entirely conceptual. There is no definitive (ideal, archetypical) "tree". Even the picture of a tree Saussure uses to represent the signified is itself just another signifier. This aside, it is Saussure's argument that it is only the consistency in the system of signs that allows communication of the concept each sign signifies.
The object itself - a real tree, in the real world - is the referent. For Saussure, the arbitrary involves not the link between the sign and its referent but that between the signifier and the signified in the interior of the sign.
In Jabberwocky, Lewis Carroll exploits the arbitrary nature of the sign in its use of nonsense words. The poem also demonstrates very clearly the concept of the sign as a two sided psychological entity, since it is impossible to read the nonsense words without assigning a possible meaning to them. We naturally assume that there is a signified to accompany the signifier.
The concepts of signifier and signified could be compared with the Freudian concepts of latent and manifest meaning. Freud was also inclined to make the assumption that signifiers and signifieds are inseparably bound. Humans tend to assume that all expressions of language mean something. In further support of the arbitrary nature of the sign, Saussure goes on to argue that if words stood for pre-existing concepts they would have exact equivalents in meaning from one language to the next and this is not so. Different languages divide up the world differently. To explain this, Saussure uses the word bœuf as an example. He cites the fact that while, in English, we have different words for the animal and the meat product: Ox and beef, in French, bœuf is used to refer to both concepts. A perception of difference between the two concepts is absent from the French vocabulary. In Saussure's view, particular words are born out of a particular society’s needs, rather than out of a need to label a pre-existing set of concepts.
But the picture is actually more complicated, through the integral notion of 'relative motivation'. Relative motivation refers to the compositionality of the linguistic system, along the lines of an immediate constituent analysis. This is to say that, at the level of langue, hierarchially nested signifiers have relatively determined signified. An obvious example is in the English number system: That is, though twenty and two might be arbitrary representations of a numerical concept, twenty-two, twenty-three etc. are constrained by those more arbitrary meanings. The tense of verbs provides another obvious example: The meaning of "kicked" is relatively motivated by the meanings of "kick-" and "-ed". But, most simply, this captures the insight that the value of a syntagm-- a system-level sentence-- is a function of the value of the signs occurring in it. It is for this reason that Leonard Bloomfield called the lexicon the set of fundamental irregularities of the language. (Note how much of the 'meaningfulness' of 'The Jabberwocky' is due to these sorts of compositional relationships!)
A further issue is onomatopoeia. Saussure recognised that his opponents could argue that with onomatopoeia there is a direct link between word and meaning, signifier and signified. However, Saussure argues that, on closer etymological investigation, onomatopoeic words can, in fact, be coincidental, evolving from non-onomatopoeic origins. The example he uses is the French and English onomatopoeic words for a dog's bark, that is Ouaf Ouaf and Bow Wow.
Finally, Saussure considers interjections and dismisses this obstacle with much the same argument i.e. the sign / signifier link is less natural than it initially appears. He invites readers to note the contrast in pain interjection in French (aie) and English (ouch).
The value of a sign is determined by all the other signs in the langue.
Saussure realized that if linguistics was going to be an actual science, language could not be a mere nomenclature; for otherwise it would be little more than a fashionable version of lexicology, constructing lists of the definitions of words. Thus he argued that the sign is ultimately determined by the other signs in the system, which delimit its meaning and possible range of use, rather than its internal sound-pattern and concept. Sheep, for example, has the same meaning as the French word mouton, but not the same value, for mouton can also be used to mean the meal lamb, whereas sheep cannot, because it has been delimited by mutton.
Langue is therefore a system of interdependent entities. But not only does it delimit a sign’s range of use, for which it is necessary, because an isolated sign could be used for absolutely anything or nothing without first being distinguished from another sign, but it is also what makes meaning possible. The set of synonyms redouter (“to dread”), craindre (“to fear”), and avoir peur (“to be afraid”), for instance, have their particular meaning so long as they exist in contrast to one another. But if two of the terms disappeared, then the remaining sign would take on their roles, become vaguer, less articulate, and lose its “extra something”, its extra meaning, because it would have nothing to distinguish it from.
This is an important fact to realize for two reasons: (A) it allows Saussure to argue that signs cannot exist in isolation, but are dependent on a system from within which they must be deduced in analysis, rather than the system itself being built up from isolated signs; and (B) he could discover grammatical facts through syntagmatic and paradigmatic analyses.
Langue works through relations of difference, then, which place signs in opposition to one another. Saussure asserted that there are only two types of relations: syntagmatic and paradigmatic. The latter is associative, and clusters signs together in the mind, producing sets: sat, mat, cat, bat, for example, or thought, think, thinking, thinker. Sets always involve a similarity, but difference is a prerequisite, otherwise none of the items would be distinguishable from one another: this would result in there being a single item, which could not constitute a set on its own. The former is temporal, and always involves at least two signs: against all, for example, or God is good. Since syntagmas can belong to speech, the linguist must identify how often they are used before he can be assured that they belong to the langue.
These two forms of relation open linguistics up to phonology, morphology, syntax and semantics. Take morphology, for example. The signs cat and cats are associated in the mind, producing an abstract paradigm of the word forms of cat. Comparing this with other paradigms of word forms, we can note that in the English language the plural often consists of little more than adding an s to the end of the word. Likewise, in syntax, through paradigmatic and syntagmatic analysis, we can discover the grammatical rules for constructing sentences: the meaning of je dois (“I should”) and dois je? (“Should I?”) differ completely simply because of word order, allowing us to note that to ask a question in French, you only have to invert the word order.
Language that is studied synchronically is "studied as a complete system at a given point in time" (The AB axis). Language studied diachronically is "studied in its historical development" (The CD axis). Saussure argues that we should be concerned with the AB axis (in addition to the CD axis, which was the focus of attention in Saussure's time), because, he says, language is "a system of pure values which are determined by nothing except the momentary arrangements of its terms". We could study chess diachronically (how the rules change through time) or synchronically (the actual rules).
To illustrate this, Saussure uses a chess metaphor. In chess, a person joining a game’s audience mid-way through requires no more information than the present layout of pieces on the board. They would not benefit from knowing how the pieces came to be arranged in this way.
A portion of Course in General Linguistics comprises Saussure’s ideas regarding the geographical branch of linguistics.
According to Saussure, the geographic study of languages deals with external, not internal, linguistics. Geographical linguistics, Saussure explains, deals primarily with the study of linguistic diversity across lands, of which there two kinds: diversity of relationship, which applies to languages assumed to be related; and absolute diversity, in which case there exists no demonstrable relationship between compared languages. Each type of diversity constitutes a unique problem, and each can be approached in a number of ways.
For example, the study of Indo-European and Chinese languages (which are non-related) benefits from comparison, of which the aim is to elucidate certain constant factors which underlie the establishment and development of any language. The other kind of variation, diversity of relationship, represents infinite possibilities for comparisons, through which it becomes clear that dialects and languages differ only in gradient terms. Of the two forms of diversity, Saussure considers diversity of relationship to be the more useful with regard to determining the essential cause of geographical diversity.
While the ideal form of geographical diversity would, according to Saussure, be the direct correspondence of different languages to different areas, the asserted reality is that secondary factors must be considered in tandem with the geographical separation of different cultures.
For Saussure, time is the primary catalyst of linguistic diversity, not distance. To illustrate his argument, Saussure considers a hypothetical population of colonists, who move from one island to another. Initially, there is no difference between the language spoken by the colonists on the new island and their homeland counterparts, in spite of the obvious geographical disconnect. Saussure thereby establishes that the study of geographical diversity is necessarily concentrated upon the effects of time on linguistic development. Taking a monoglot community as his model (that is, a community which speaks only one language), Saussure outlines the manner in which a language might develop and gradually undergo subdivision into distinct dialects.
Saussure’s model of differentiation has 2 basic principles: (1) that linguistic evolution occurs through successive changes made to specific linguistic elements; and (2) that these changes each belong to a specific area, which they affect either wholly or partially.
It then follows from these principles that dialects have no natural boundary, since at any geographical point a particular language is undergoing some change. At best, they are defined by “waves of innovation”—in other words, areas where some set of innovations converge and overlap.
The “wave” concept is integral to Saussure’s model of geographical linguistics—it describes the gradient manner in which dialects develop. Linguistic waves, according to Saussure, are influenced by two opposed forces: parochialism, which is the basic tendency of a population to preserve its language’s traditions; and intercourse, in which communication between people of different areas necessitates the need for cross-language compromise and standardization. Intercourse can prevent dialectical fragmentation by suppressing linguistic innovations; it can also propagate innovations throughout an area encompassing different populations. Either way, the ultimate effect of intercourse is unification of languages. Saussure remarks that there is no barrier to intercourse where only gradual linguistic transitions occur.
Having outlined this monoglot model of linguistic diversity, which illustrates that languages in any one area are undergoing perpetual and nonuniform variation, Saussure turns to languages developing in two separate areas.
In the case of segregated development, Saussure draws a distinction between cases of contact and cases of isolation. In the latter, commonalities may initially exist, but any new features developed will not be propagated between the two languages. Nevertheless, differentiation will continue in each area, leading to the formation of distinct linguistic branches within a particular family.
The relations characterizing languages in contact are in stark contrast to the relations of languages in isolation. Here, commonalities and differences continually propagate to one another—thus, even those languages that are not part of the same family will manage to develop common features.