Arbitrariness and Rationality

Vít Gvoždiak

The Institute of Philosophy of the Academy of Sciences Prague, Czech Republic

Abstract

Keywords: semiology, linguistic sign, Ferdinand de Saussure, convention, motivation

Everything that relates to language as a system must, I am convinced, be approached from this viewpoint, which has scarcely received the attention of linguists: the limiting of arbitrariness. This is the best possible basis for approaching the study of language as a system. In fact, the whole system of language is based on the irrational principle of the arbitrariness of the sign, which would lead to the worst sort of complication if applied without restriction. (CGL, p. 133)

This statement from the sixth chapter of the second part of theCourse in General Linguisticsmade in the context of explaining the relation between absolute arbitrariness and relative arbitrariness, points out that language systems do not operate in harmonic accordance with the arbitrary nature of linguistic signs and hints at the former being only possible as a restriction imposed on the latter.

In case of arbitrariness—as with many other Saussure’s terms—there are two types of false preconceptions. The first results mostly from the notoriously problematic posthumous edition of theCourse, never authorized by de Saussure himself, which in many respects poses a philological challenge. For this reason, I will try to broaden the resource pool as much as possible, and in addition to the 1916 version of theCourse in General Linguistics(hereinafter referred to asCGL) as edited by C. Bally and A.Sechehay will also draw on student notes from Saussure’s three lecture cycles (namely the second and third cycle/volume, as these are relevant to the present paper; I refer to them as Saussure 1997 and Saussure 1993) as well as Saussure’s manuscript found in 1996 and published in French in 2002 asÉcrits de linguistique générale(which I refer to in its English translation from 2006 asWGL). Since my goal is not to reconstruct the historical genesis of the concept of arbitrariness but rather use these sources as inputs for more systematic considerations, the second type of preconception to be dispelled is that arbitrariness—even though it is called the first principle and one of the basic principles of any type of semiological study (CGL, p. 67)—cannot be assigned the status of a singular linguistic principle. The notion that a language phenomenon can be explained by means of a single, unified methodological principle is wrong for the same reason the notion that language itself is a single, unitary entity is mistaken (WGL, p. 161), therefore, plurality nature of approaches is a principal precondition of all linguistic investigations (WGL, p. 5). More precisely, as formulated in theCGL: when examining a language, the object is not givena priori, but “it is the viewpoint that creates the object” (CGL, p. 8). Saussure’s viewpoint is formed by a set of well-known mechanisms (sign, value, langue, etc.)—each of which in itself represents an important cornerstone of structural semiology. These mechanisms are nevertheless at the same time interconnected and intertwined and their unravelling presents a significant challenge. What follows is an attempt to answer the questions of how exactly the differentiation between absolute and relative arbitrariness in Saussurean semiology relates to the systemic nature of language, what constitutes the irrationality of arbitrariness, and what could be considered its rational counterpart.

1. Arbitrariness and Relative Arbitrariness

With respect to the linguistic sign, arbitrariness is most often considered to be its first principle (CGL, pp. 67-70; Saussure, 1993, p. 87). It represents not only an immediate continuation of the ideas of Dwight Whitney (Koerner, 1973, p. 315), which Saussure was well familiar with (see e.g. Jakobson, 1971), but also a contribution to the complex discussion that dates all the way back to classical antiquity, as personified by Cratylus and Hermogenes, and which often led modern expounders of Saussure’s ideas to misleading interpretations, especially in relation to Peirce’s semeiotic.

Among the several formulations of the problem offered by theCGLand other Saussure’s texts, we can for simplicity’s sake pick the one below listing the two fundamental features of arbitrariness (both have resulted in a number of critical responses, among the best known of which are Bolinger, 1965; Jakobson, 1962,1966; Benveniste, 1971):

(1) Anti-reductiveness: even though it is, within a sign, possible to differentiate between the signifying and the signified component (signifiantandsignifié), their unification in the unity of the sign is not determined by anything outside of the sign.Thesignifiant-signifiépair at the same time cannot be reduced to a single unit, either mental or physical.

(2) Stabilization and immanence: arbitrariness works as the stabilization principle of signs with respect to the mutable, individual psychology of their users and mutability of the extra-sign world. Therefore, Saussure identifies arbitrariness with unmotivatedness (Saussure, 1993, p. 85). Linguistic sign as a combination ofsignifiantandsignifiéis not determined by either purposes and intentions of speakers, nor by the nature of what the speakers talk about; arbitrariness is therefore an important cornerstone of Saussurean immanentism. A sign is seen as in no way connected to the thing it refers to. Constitution of a sign is undetermined by reference; in other words,signs are “free and arbitrary in relation to the object” (WGL, p. 140).

It would then appear, at least when considering the genesis of theCGLthat arbitrariness was not initially seen by Saussure as particularly problematic. Indeed, he did not handle arbitrariness in his first lecture cycle at all. A brief mention appears in the second cycle: basic properties of writing (which are essentially similar to those of language in general) include the arbitrary nature of signs, and therefore nonexistence of (causal) relation between signs and the things they refer to; the character of sign value is purely that of differentiation, and therefore negative (Saussure, 1997, pp. 7,113; cf.WGL, p. 147). Arbitrariness is given closer attention in the last lecture cycle,where it is seen as the most important property of language and as such plays a key role in semiology (Saussure, 1993, p. 76; cf. Koerner, 1973, p. 335).

Even though language can be considered an institution of sorts, it is precisely its arbitrariness that makes it idiosyncratic and incomparable to other human institutions(Saussure, 2006, pp. 146-147). In case of customs, laws, fashion or symbols, we are dealing with a “necessity” between means and ends, and “the natural relation of things” (CGL, p. 75;WGL, p. 147). Language on the other hand is autonomous in selection of its means and its identification with other institutions only leads to misunderstandings (WGL, p. 149). Even though arbitrariness serves as a functional distinctive feature on the level of institutions (sign systems), it cannot be seen as a feature which exhausts the properties of the system of language. On the contrary, the notion of an entirely arbitrary system of signs is impossible. Linguistic arbitrariness/unmotivatedness cannot be reduced to zero (Saussure, 1993, p. 89), while the opposite pole of complete motivatedness is equally impossible. Absolute motivatedness and absolute arbitrariness are constructs that refer to abstract limits of possibility and which in their pure form do not exist in any language (CGL, p. 133).

Arbitrariness in language is, in general, limited by solidarity with the tradition of the language’s past states (CGL, pp. 73-74), i.e. by its temporality. If language could be a non-temporal system, it most likely could and would also be completely arbitrary. It is precisely temporality that prevents complete freedom of selection in language. The second limiting principle is the differentiating nature of language, or its synchronic systemic character, which prevents revolutionary changes and complete randomness. Saussure and theCGLattempt to weaken the power of arbitrariness by means of introducing the concept of relative arbitrariness.

The difference between absolute and relative arbitrariness is discussed only in the third lecture cycle (Saussure, 1993, pp. 85-90), in relation to the question of abstract units. In theCGL(p. 132), this appears as continuation of the differentiation between associative and (especially) syntagmatic relations. From the associative (paradigmatic)standpoint, the expression “dix-neuf” (nineteen) is motivated by its relations to expressions such as “dix-huit” (eighteen) or “soixante-dix” (seventy); from the syntagmatic standpoint, it is motivated by its components “dix” and “neuf” (CGL,p. 131). Both systemic relations thus at first glance seem to neutralize the proposed absolute arbitrariness of linguistic signs; on closer inspection we nevertheless see that the components of a sign motivated in this way are themselves arbitrary and,moreover, the differentiating value of the whole motivated sign is not a simple sum (of values) of its components (CGL, p. 132).

Despite the suggested terminological similarity and the mutual conceptual relatedness, we need to distinguish between absolute and relative arbitrariness, based mostly on the following pair of features:

(1) Absolute arbitrariness and relative arbitrariness (or motivatedness) are not principles of the same order. Relative arbitrariness is a more complex systemic relation. It has a dual nature; it includes the internal relation between signifiantandsignifiéas well as the external relation between terms as complete signs. Relations between terms (signs) are in all probability secondary—provided that no two signs exist such that each of them separately consists of a combination ofsignifiantandsignifié, we cannot consider external relations at all. Internal relations may be independent of external relations, whereas each external relation does presuppose(at least two) internal relations (Saussure, 1993, p. 90). Absolute arbitrariness is a precondition of relative arbitrariness/motivatedness. Further terminological and conceptual support for the belief in this categorical difference rests in the fact that relative arbitrariness and relative motivatedness—due precisely to their relativity—may function as synonyms, or, to use Saussurean diction, represent two sides of the same coin. There is, however, no such synonymy with absolute arbitrariness; absolute arbitrariness and absolute motivatedness are the extreme poles of an opposition.

(2) Absolute arbitrariness cannot constitute the foundations of sign value or language system. The language system,langue, is dependent on intra-systemic differences (WGL, p. 189), it is a set of values, “the existence of which is only made possible due to the fact of their being in opposition” (WGL, p. 51). From the standpoint of both modalities of arbitrariness, this fact is formulated relatively clearly:“Reduction in any system of langue of absolute arbitrariness to relative arbitrariness;this is what makes up the ‘system’.” (WGL, p. 233) The third lecture cycle makes it apparent exactly how important Saussure considered the fact that without the internal relation of arbitrariness, two words (two signs) cannot be compared with respect to their relative motivatedness. A sign is built on an autonomous internal relation betweensignifiantandsignifiéwhich is not or does not have to be affected by its relation to other signs at all. We could say that absolute arbitrariness involves a certain type of atomism, something Saussure however was most likely aware of,when he referred to the arbitrariness of linguistic signs as an “irrational principle”, the limitation of which is a precondition of the study of language as a system (Saussure,1993, p. 87). It is in fact the external relation of relative motivatedness that represents one of the principles of order and regularity (CGL, p. 139) and aids in construction of language system in the narrower sense by reduction of absolute arbitrariness to relative arbitrariness, that is to say, a “relatively rational” language system, which can be studied by logic and grammar (Saussure, 1993, p. 96), or even studied at all.

伊顿是一家全球领先的动力管理公司，2016年销售额达197亿美元。伊顿提供各种节能高效的解决方案，以帮助客户更有效、更安全、更具可持续性地管理电力、流体动力和机械动力。伊顿致力于利用动力管理技术和服务，提高人类生活品质和环境质量。伊顿在全球拥有约9.5万名员工，产品销往超过175个国家和地区。

2. Rationality and Irrationality

Saussure does not give any special, explicitly defined attention to rationality, it nevertheless does appear in several places in his writings as an attribute which is supposed to associate correctness of explanation or of a certain categorization. For instance, differentiation of the object of linguistic study into speaking and language which is in turn further divided into synchrony and diachrony is seen as “the rational form” (CGL, p. 98). Elsewhere, Saussure takes this even further when he asserts that even though linguistic facts must be investigated by different methods (WGL,p. 5), due to the nature of language linguistics is nevertheless obligated to use only two standpoints, synchronic and metachronic, i.e. diachronic (WGL, p. 188).TheCGLchapter “Rational Divisions” provides a list of categories of subjects of linguistic disciplines—i.e. morphology, lexicology, syntax—based on associative and syntagmatic relations. Study of a living language has to employ the rational method which “consists of (a) setting up the system of sounds as revealed by direct observation, and (b) observing the system of signs used to represent—imperfectly—these sounds” (CGL, p. 37). The completely non-systematic use of the term shows that the “rational” attribute is used in at least two non-synonymous ways:

(1) Rational, meaning motivated. A relation is rational given that it is causal or has some sort of its own norm. Given this standpoint, non-arbitrary and symbolic systems may be seen as rational because sign units of such systems have “rational” relations to the things that are being referred to (CGL, pp. 68, 73). Monogamous matrimony is (probably) more reasonable than polygamy, whereas the sign relation between the expression “cow” and the notion of cow is based on “pure unreason” (WGL, p. 149).The form of matrimony and the suitability of any of its variations can be subject to rational or philosophical discussion (WGL, p. 149;CGL, p. 73) while an arbitrary sign cannot, because “any subject in order to be discussed must have a reasonable basis” (CGL, p. 73). If, however, language unlike all other human institutions is in fact completely arbitrary, not based on any reasonable norm, and its speakers use it without having explicit knowledge of its system, this then also means that arbitrariness is the cause due to which language itself is principally beyond rational and scientific, linguistic understanding. It is along these lines that one could probably interpret the complaint (CGL, p. 133) that language is not fully rational. If it were, it would be possible to study it “independently”.

It is, however, precisely the combination of the notions of “independently”and “rational” which in this case leads to a dead end. This is because Saussurean semiology uses the conception oflangueand attempts to propose study of language in itself, its special character determined by the arbitrary nature of signs. “Full rationality”—if we were to understand it in the sense of “full motivatedness” or “full symbolicity”—however upsets this notion of an independent language which was as an ideal form supposed to be the subject of interest of linguistics. In other words,should language be fully rational, we could not study it separately and independently because its units (signs) would be causally linked to the things they refer to. At the same time, if we were to study language as a completely independent phenomenon, its study could not be rational.

Saussure rules out these extremes not by opting for one of the absolutes and refusing the other, but rather by reference to “relative rationality” which governs language, yet at the same time allows for it to qualify as an object which can be studied reflectively, i.e. rationally (CGL, p. 73). The attributes of “rational” and“irrational” then do not amount to much more than abstract conceptions which,similarly to “absolute motivatedness” and “absolute arbitrariness”, do not set the rules of any specific games, or an even more specific game strategy, but rather attempt to outline the conditions for existence of playgrounds and spaces for games to be played in, regardless of the actual games being played there and whose arbitrariness and rationality at each given moment are necessarily relative. If we were to understand arbitrariness and rationality in one of the specific forms, Saussurean semiology would find itself in the unenviable position of trying to avoid its own redundancy on the one hand and actual unfeasibility as a scientific discipline on the other hand.

(2) Rational, meaning conventional. In contrast to the aforementioned identification of the rational with the motivated, there is a somewhat cryptic statement in theCGLto the effect that given precisely the arbitrariness of linguistic signs, language system appears to be organized “freely” based on “a rational principle” (CGL, p. 78).

The statement can nevertheless also be understood as saying that the arbitrary character of the units is not based on the exclusion that is seemingly hinted at, but rather in a curious connection between the “free” and the “rational” which cannot be reduced to rational motivatedness. This “free rationality” could simply be called conventionality and identified with arbitrariness. This is supported by Saussure’s familiarity with the works of Dwight Whitney, who conceived of “arbitrary” and“conventional” as synonyms (Koerner, 1973, p. 317; cf.CGL, pp. 10, 76), as well as certain statements by Saussure which seem to hint at the fact that the life of signs is based on a contract, without which a sign cannot exist (WGL, p. 68). Such understanding of “arbitrary convention” is most probably given its most explicit formulation in theCGL’s section on immutability and mutability of the sign (CGL,pp. 71-78), which points out thede factoinseparable connection of language with the community of speakers and time. The hypothetical rational principle which organizes language is not in contradiction of the social character of language, because it does not work based on a purely “logical base” (in this case, the formulations of theCGLsuggest that the reader identifies “rational” with “logical”), but it is in contradiction of the social force which works within time. The social force in time is the principle of continuity which “cancels freedom” (CGL, p. 78), and, should we understand it as a different expression for “convention” (“free” in the sense of “dynamic”), we would at the same time have to refuse its identification with arbitrariness, because it is not(absolute) arbitrariness but rather continuity which forms the basis of sign connections betweensignifiantandsignifié.

All these interpretations are evidently faced with a number of difficulties, and there is unfortunately no straightforward path to solving them. Two more general facts can nevertheless be taken into consideration, which are referred to repeatedly by Saussure in his texts and which lead to the belief that, from the standpoint of rationality, internal or absolute arbitrariness represents a sort of threshold for semiological inquiry, below which Saussurean semiology cannot operate but without which it at the same time cannot exist.

The first of these is the assertion (CGL, p. 122) that language has at its disposal only complex units whose complexity is “deprived of any natural unity” and which cannot be further simplified. The basic language units are not to be thought of as chemical elements, but rather as chemical mixtures (WGL, p. 4). Searching for rational relations below the level of these mixtures is meaningless in the same way as it is meaningless to look for unified language facts independently with respect to phonological and psychological facts, albeit they may in fact exist that way(WGL, p. 68). If the systematicity oflangueis from the standpoint of arbitrariness a secondary property, then it must be acknowledged that rational relations are likewise qualitatively secondary and that no decompositional analysis of sign can either capture or isolate them. An analytical approach may with respect to language facts identify a pair of sign components/objects,signifiantandsignifié, or—as they are called elsewhere, using more obscure terminology—signs and ideas, these however do not constitute objects of the same kind and any independent considerations of them as separate entities only lead to postulation of “two different grammars” (the grammar of ideas and the grammar of signs), each of which is “incorrect and incomplete” (WGL,p. 5). Given this standpoint, there is no rationality in language such as would not be a rationality of complex units.

The other fact is the concept of differentiating value, probably the most important cornerstone of systematic semiology which at the same time plays a key role in differentiating between absolute and relative arbitrariness. The sign value which—in surprising opposition to theCGL(p. 114)—Saussure in his manuscript principally identifies with meaning, significance, function and even use (WGL, p. 12) cannot be considered only from the standpoint of the internal, absolutely arbitrary relation betweensignifiantandsignifié.

Saussure brings up this fact repeatedly. One comprehensive example being for instance the following set of assumptions which Saussure proposed in his manuscript(WGL, p. 20): (1) The sign exists solely based on its meaning, (2) meaning exists solely based on its sign, (3) signs and meanings exist solely based on the difference between signs. The first two assumptions represent a mirror requirement of the obligatory reciprocity ofsignifiantandsignifié. The third assumption then represents a conglomerate of two levels of difference, on the one hand a purely negative level,which articulatessignifiantandsignifiéseparately, on the other hand a positive level of its kind, which operates at the level of signs as complex units. The first of these levels—the comparison of the units ofsignifiantand the units ofsignifiéindividually—is a difference in the proper sense for its pure negativity and, above all, because it delimits both components of the sign which have no existence of their own before this differentiation. Relations at the second level should not be thought of as relations of difference, but rather as relations of units which, because of their positive nature, are only distinct (CGL, p. 121). Arbitrariness seems to be in some way breaking between these the two levels. From the point of view of thesignifiantsignifiérelation, arbitrariness assumes existence of negative values of the sign components, i.e. differential values in the proper sense, corresponding to the first level. At the same time, arbitrary relation of sign components is itself a prerequisite for determining the value of the sign as a whole, which corresponds to the second level. Even though Saussure (WGL, p. 238) in relation to definition of semiology as a discipline concerned with values which are not connected to extra-linguistic “things”by any sort of causal or rational relations speaks of “arbitrarily determinable value”,arbitrariness does not in fact construct the system in any way, nor can it principally do so, as this would lead to “the worst sort of complication”.

Since “[a]rbitrary and differential are two correlative qualities” (CGL, p. 118),arbitrariness is a necessary part of the sign value delimitation mechanism; at the same time, it requires the first level and conditions the second. Its primary use for Saussure is however more that of a shorthand answer to those who would insist to pollute the immanent “purity” of language by extra-linguistic motivatedness. Should—hypothetically—components of signs be unified based on some sort of external or rational motivation, then (linguistic) values would also be determined by external motivation and necessarily precede the system, precede linguistic differentiation. The irrationality of absolute arbitrariness is a precondition of relative sign values, because it is a precondition of the very existence of an autonomous language system which—should it be motivated/rational—would be a thoroughly non-autonomous system and could only be studied in the same manner as other symbolic systems. The very notion of the language system, and, by the same token, linguistics and semiology as independent scientific disciplines would in that case appear superfluous.

3. Conclusion

Even though the terms rational and rationalistic are not synonymous and the way Saussure uses the former is not systematic, while he does not use the latter at all,Valentin Voloshinov suggests considering Saussure’s semiology as a rationalistic project. His main reason for this (Voloshinov, 1973, pp. 57-58) is the fact that arbitrariness and conventionality of language is a characteristic property of philosophical rationalism of the 18th century, which, being a study of forms, abstracts from ideological contents and focuses solely on “the inner logic of the system of signs itself” (Voloshinov, 1973, p. 58). It is therefore not difficult for Voloshinov to define Saussure’s semiology precisely as based on the conception of arbitrariness, with the Leibnizian idea of universal grammar. Although certain support for this notion could be found e.g. in Saussure’s assertion that linguistics cannot consider sensory data to be the source of its inquiries (WGL, p. 5), and that (linguistic) sign is not the opposite of the physical and the mental, because “[t]here is one domain, interior,psychic, where both sign and meaning are to be found, bound indissolubly one to the other” (WGL, p. 6), this forms a rather vague support, against which there could be raised arguments at the very least from the standpoint of its problematic conception of substance. Voloshinov’s line of thought nevertheless gives, albeit indirectly, rise to two questions worth attention. The first of these is the relation between (linguistic)sign and (language) system which finds its expression in assertions on “language as a system of arbitrary signs”; the other is then the relation between arbitrariness and convention, which finds its expression in assertions on “arbitrary convention”.

The relation between sign and system is primarily a question of the mechanism of value, without which it would not be possible to consider the systematic nature of language at all (CGL, p. 112). The role played by arbitrariness here may therefore seem vague or even inappropriate. In his critique of arbitrariness, Émile Benveniste(1971) says that the union of sign components is not arbitrary, but rather necessary;Saussurean semiology based on arbitrariness is then not a manifestation of rationalism,but rather of philosophical relativism. The assertion that “the whole system of language is based on the irrational principle of the arbitrariness of the sign” (CGL,p. 133) cannot be understood as a belief in arbitrariness as a constituting principle of the system of language. It can only be understood as a constitutive principle in the sense of a force in opposition to which linguistic system is formed. It therefore does not come as a surprise that absolute arbitrariness is seen by many commentators as“an excessive relativist rigidity that contradicted the general ‘spirit’ of Saussure’s teaching” (Gasparov, 2012, p. 73), because the two most important concepts—those of arbitrary sign and the system—appear as opposing forces.

When Saussure states that “language represents arbitrary convention” (WGL, p.140), he is in fact guilty of a certain simplification for the sake of explanation—a similar formulation is used by Voloshinov (1973, p. 58) when he speaks of “language as a system of conventional, arbitrary signs”—and with this simple gesture identifies two principles which are correlated but hardly identical. Arbitrariness can be seen as a pre-rational precondition of constitution of a sign or a linguistic system whicha priorirules out any determination other than an internal linguistic one, without however constituting a construction mechanism of the said determination. Convention in the sense in which it is used by Saussure is largely of a pragmatic nature; it does not precede the system as its precondition but is rather formed—as acknowledged by theCGL(p. 78), after all—continually (see mainlyWGL, p. 67) as a combination of the social force and temporal nature of language. It is therefore superfluous at best to speak of “arbitrary convention”, because every convention presupposes arbitrariness.Every convention—as accentuated by David Lewis (2002, p. 70)—is a convention due to, among other things, the principal necessity of existence of at least one alternative which could fulfill the role of the existing convention with the same effectiveness.Should there be no such alternative, there would be no convention either, but rather an unavoidable necessity, that is to say, a necessity which could be found in symbolic systems with their own, non-arbitrary bases. Arbitrariness is, given this standpoint,a plurality of possibilities, which can be entirely irrational, because the number of options within complex systems is infinite (seeWGL, p. 52). Convention, on the other hand, is selection of a specific option over others. While from the standpoint of the continuity principle, selection as convention (CGL, p. 78) is not entirely rational,from the standpoint of cultural dynamics it represents a necessary precondition for the establishment of convention. This selection is not and in principle cannot be a conscious decision of any particular individual, but only a collective, cultural custom or pattern (CGL, p. 68). This leads to somewhat indeterminate conclusion that selection is not only a prerequisite, but also a product of the dynamics of interpersonal linguistic dynamics.