As a contribution to the discipline of semiology established by Ferdinand de Saussure in his Course in General Linguistics (1916), Louis Hjelmslev’s Prolegomena to a Theory of Language (1943) is both provocative and problematic. Taking from Saussure his two essential axioms, that “the true and unique object of linguistics is language studied in and for itself,” and that “language is a form and not a substance” (Saussure 232, 122), Hjelmslev proceeds to analyze language as a “self-sufficient totality, a structure sui generis” (Hjelmslev 6), aiming for a definition of his “object by an arbitrary and appropriate strategy of premisses [sic]” (15). This means that the terms of his definition will be internal to language, derived from it, and thus unmotivated by the systems and methods of other disciplines. His singular methodological precommitment is to the “empirical principle”—that “description shall be free of contradiction (self-consistent), exhaustive, and as simple as possible” (11). By adhering to this principle, Hjelmslev intends to establish linguistics as a “systematic, exact, and generalizing science” (6, 9). But in taking this approach to the science of language, Hjelmslev must wrestle with the tensions produced by the application of deductive empiricism to a formal whole. If Saussure’s revolution was in the refutation of substantialist linguistics (Stein 6), then it remains to be seen whether Hjelmslev’s methodology will lead him to a regressive position, or a true advancement of his predecessor’s thought.
By virtue of its style, and one particular quirk of Hjelmslev’s writing, the Prolegomena can be quite difficult to disentangle. Stylistically, it is neither a course, nor a systematic treatise, but a theory, a long essay that develops its system as it unfolds. Unlike Saussure’s semiology, which can be summarized by a single principle—the “reciprocating” or differential “mechanism of language” (Stein 6)—Hjelmslev constantly deals in “operative definitions,” repeatedly changing his terminology over the course of the text (Hjelmslev 21). One must be diligent in tracking the multiple terminological equivalences that he establishes, and the subtle nuances between them, so as not to become lost in the jargon—especially if one is to identify the points of continuity between the Prolegomena and Saussure’s Course. To this end, a brief repetition of Saussure’s framework will provide us with a firm basis from which to discuss Hjelmslev’s theory.
First, for Saussure, the goal of semiology is to discover in speech (parole) that which is “common” in “all other semiological systems”—the structure, principle, or mechanism of signification itself: language (langue, and not langage) (17). Both parole and langue are language, but it is langue which is to him of chief interest. Second, the sign is an association of “two terms,” the “signified” or “concept” with the “signifier” or “sound-image” (65-67). The sign is a dual “entity,” irreducible to either of its terms (66). Third, the sign has “two primary characteristics”: arbitrariness and linearity. Arbitrariness means that the relations between signifier and signified, and sign and referent, are purely “conventional”; linearity means that signs are “presented in succession; they form a chain” (67, 69, 70). Fourth, because the sign is arbitrary, it can also be characterized as “differential”—“Signs function . . . not through their intrinsic value but through their relative position” (118). The sign has no substance—it is a form, a function, a relation. Fifth, in that langue is a “system of signs,” langue is therefore a “system for equating things of different orders”—precisely, concepts and sound-images (15, 79). In this, it is necessary to maintain a constant “distinction between the system of values per se and the same values as they relate to time.” This distinction Saussure schematizes as the axes of “simultaneities” and “successions” (80). The axis of simultaneities, concerning the “system of values per se,” corresponds to the discipline of “synchronic” or “static” linguistics, and the axis of successions, concerning the “values as they relate to time,” to “diachronic” or “evolutionary” linguistics (81). Sixth, because the sign is differential, and language as a system of signs is thus a system of relations or differences (120), the “mechanism of language” is a “reciprocating function” operating between two primary relations: the “associative” (or paradigmatic) and the “syntagmatic” (123). Thus, from the smallest unit of the sign system, the dual entity of the sign, to the sign system as a whole, the language, we see Saussure’s axial structure recapitulated, and an equation of terms established: langue = simultaneity = synchrony = association = paradigm and parole = succession = diachrony = syntagm. At the level of the individual sign, we see a sign like défaire (Saussure’s example) in both paradigmatic relation to other signs (by the simultaneity of root paradigms—DÉ-coller, contre-FAIRE—and phonemic paradigms: d or t, é or u), and in internal syntagmatic relation (by the succession of roots—dé + faire—and phonemes—d + e + f + ɛ + ʁ). At the top level, we see an entire sign system, like the English language, as it operates at present (the simultaneity of its myriad associative relations), and as it has evolved through time, from Old English to Middle English to Modern English (the succession of its synchronic states) (129). In sum, it is plain to see that Saussure’s framework is consistent throughout—the differential function of the linguistic mechanism is at work at every level.
How does Hjelmslev compare? For Hjelmslev, language is both a “process” and a “system” (9). This division corresponds directly to Saussure’s division of speech and language, parole and langue. Through an analysis of “processes or texts,” Hjelmslev aims to deduce the “system or language on which all texts of the same premised nature are constructed” (16). This, too, corresponds to Saussure’s derivation of the linguistic mechanism (langue) from everyday speech (parole). Also for Hjelmslev, the system or “totality does not consist of things but of relationships,” and so the “priority of dependences in language” must be “recognized” (23). Like in Saussure, language, for Hjelmslev, is a system of signs. In the Prolegomena Hjelmslev’s treatment of language follows Saussure’s definition of it as a form “based entirely on the opposition of its concrete units” (Saussure 107)—those units being the “pure values [i.e., relations] . . . determined by nothing except . . . momentary arrangement” or position (Saussure 80)—indeed, the “interdependent whole” that is “without positive terms” (Saussure 113, 120). Hjelmslev makes clear that linguistic science must be wary of falling into “naive realism,” which “suppose[s] that analysis consist[s] merely in dividing a given object into parts.” On the contrary, the “important thing is not the division of an object into parts, but the conduct of analysis so that it conforms to the mutual dependences between these parts” (Hjelmslev 22). Indeed, language has
existence only by virtue of these dependences; the whole of the object under examination can be defined only by their sum total; and each of its parts can be defined only by the dependences joining it to other coordinated parts, to the whole, and to its parts of the next degree, and by the sum of the dependences that these parts of the next degree contract with each other. (23)
As such, Hjelmslev contends that the “objects” of naïve realism are . . . nothing but intersections of bundles of such dependences” (23). These “intersections” must not be substantialized; they can be given only “strictly formal” definitions (20).
Nevertheless, the dependences and intersections of language must be made “explicit” (20). Such a task requires the “registering” of “certain dependences between certain terminals,” or intersections, with the intent of delimiting a “limited number of elements recurring in various combinations” (28, 9). Hjelmslev identifies three such categories of dependence: “interdependences,” “determinations,” and “constellations.” Interdependence is a “mutual” dependence, determination is a “unilateral” dependence, and constellation is a “freer” dependence, an independent relation (24). With respect to the axis of language in question, specific terms for each dependence can be employed: in a linguistic process, or the syntagmatic axis of language, interdependence is “solidarity,” determination is “selection,” and constellation is “combination”; in a linguistic system, or the paradigmatic axis of language, interdependence is “complementarity,” determination is “specification,” and constellation is “autonom[y]” (25).
Rather than use Saussure’s terminology of axes, however, Hjelmslev refers to the linguistic process and linguistic system as “hierarchies” (29). This is to indicate the recursively propagated character of the linguistic mechanism (i.e., the reciprocating movement between the syntagmatic and paradigmatic axes). In a linguistic process, “parts” are combined into “chains,” while in a linguistic system, “members” are combined into “paradigms” (29). In the linguistic mechanism generally, “parts” and “members” are referred to as “components,” and “chains” and “paradigms” as “class[es]” (29). But, to avoid losing sight of the dependences between components, Hjelmslev prefers to use the language of “function[s]” and “functive[s]”; parts, members, and components thus become functives that serve as terminals for functions (33). What is more, due to the hierarchical structure of language, functions can also serve as functives in other functional relations. In the example from Saussure’s Course above, the roots dé + faire are functives with a syntagmatic function to each other, so acting as parts of a syntagm or chain. But dé- and -faire are also members of root paradigms, and so are functives with paradigmatic function within the syntagm. Further, as roots, dé- and -faire are of a different class from the sign défaire, and therefore are its components; the function “contracted” therein is also paradigmatic. We could speak further of phonemes syntagmatically contracted into root chains, and signs syntagmatically contracted into sentences—in each case, the paradigmatic function is always also at work, distinguishing phonemes from other members of the phonemic paradigm, sign from the sign paradigm, and so on.
Hjelmslev’s discussion of functions allows him to introduce the terminology of “conjunction” and “disjunction.” Conjunction is the “both-and” or “coexistence” function, while disjunction is the “either-or” or “alternation” function (36). Given the foregoing discussion, we can see that conjunction corresponds to the syntagmatic function, while disjunction corresponds to the paradigmatic function: in défaire the roots dé- and -faire are “conjuncts” or “coexisents,” dependant on each other in succession; similarly, the roots dé-, re-, and contre-, and –faire, -coller, -placer, are “disjuncts” or “alternants,” dependent on each other in simultaneity (37). To clarify the dependences herein, the conjunctive or syntagmatic dependence is that of “relation,” and the disjunctive or paradigmatic dependence is “correlation.” The hierarchies of process and system, parole and langue, are thus, respectively, a “relational hierarchy” and a “correlational hierarchy” (39). Thus, we can refer to solidarity, selection, and combination as syntagmatic or relational functions, and complementarity, specification, and autonomy as paradigmatic or correlational functions.
Hjelmslev does not stop here, however. If functions can be functives, what of functives that are not functions? Any such functive is an “entity,” and entities can be either “constant” or “variable”—a functive is a constant if its “presence is a necessary condition for the presence of the functive to which it has function,” and it is a variable if its “presence is not a necessary condition for the presence of the functive to which it has function” (35). The three types of dependences, as functions contracted by their functives, can be nuanced accordingly: an interdependence is “a function between two constants”; a determination is “a function between a constant and a variable”; and a constellation is “a function between two variables” (35). By extension, a solidarity is a syntagmatic function between two constants, a selection is a syntagmatic function between a constant and a variable, a combination is a syntagmatic function between two variables, a complementarity is a paradigmatic function between two constants, a specification is a paradigmatic function between a constant and a variable, and an autonomy is a paradigmatic function between two variables. Hjelmslev further organizes these functions with the categories of “cohesion” and “reciprocity” (35). Cohesions are functions contracted by “one or more constants” (interdependence and determination), while reciprocities are functions contracted by “one and only one kind” of functive (interdependence and constellation). Curiously, Hjelmslev neglects to include a function contracted by one or more variable (which would include determination and constellation). Regardless, Hjelmslev considers his description of the “self-sufficient totality” of language to be, at this point, complete, as represented in the following table (6, 43):
To Saussure’s equations, then, we could add langue = system = disjunction = alternation = correlation = equivalence and parole = process = conjunction = coexistence = relation = connection. So far, Hjelmslev theory has only extended Saussure’s system of signs.
But what of the sign itself? “That a language is a system of signs seems a priori an evident and fundamental proposition”—in the preceding analysis, we have seen Hjelmslev establish the “system” portion of the sign system, which is to say, its relational or differential form. But for “linguistic theory” to be complete, it must “tell us what meaning can be attributed . . . to the word sign” (43). The sign, or in Hjelmslev’s preferred terminology, the “sign-expression,” is “characterized . . . by being a sign for something else” (43). The sign is a “function,” and its function is to be “the bearer of meaning” (43). But if we follow Hjelmslev’s program of analysis, we cannot stop with the sign—it is not an irreducible functive, because signs, in turn, are composed of phonemes, which cannot “be said to be bearers of meaning and thus no longer are sign-expressions” (45). How can language be a system of signs, if the signs of which it is composed are not in themselves signs? This is Hjelmslev’s problematic. He concludes that linguistics must
abandon the attempt to analyze into “signs,” and . . . recognize that a description in accordance with our principles must analyze content and expression separately, with each of the two analyses eventually yielding a restricted number of entities, which are not necessarily susceptible of one-to-one matching with entities in the opposite plane. (46)
A system of dependences must be composed of a “limited number of elements,” that is, of “non-signs whose number is restricted, and, preferably, severely restricted” (9, 46). The elements or non-signs here referred to, which “enter into a sign system as parts of signs,” Hjelmslev designates “figuræ” (46).
The sign, for Hjelmslev, is a function, and so, must be decomposable into functives. Herein lies Hjelmslev’s first significant departure from Saussure. For Saussure, the sign or “linguistic unit” is a “double entity, one formed by the associating of two terms” (Saussure 65). It is a “two-sided psychological entity” in which the “two elements,” the “concept” (signified) and the “sound-image” (signifier), are “intimately united” (66). Though the two elements can be spoken of as distinct, the link between them is indissoluble. Hjelmslev would agree, in part, similarly stating that
there is . . . solidarity [syntagmatic function between two constants] between the sign function and its two functives, expression [signifier] and content [signified]. There will never be a sign function without the simultaneous presence of both these functives; and an expression and its content, or a content and its expression, will never appear together without the sign function’s also being present between them. The sign function is in itself a solidarity. Expression and content are solidary—they necessarily presuppose each other. An expression is expression only by virtue of being an expression of a content, and a content is content only by virtue of being a content of an expression. Therefore—except by an artificial isolation—there can be no content without an expression, or expressionless content; neither can there be an expression without a content, or content-less expression. (48)
The subtle difference is that, for Hjelmslev, expression and content, signifier and signified, can, in fact, appear separately as unformed sound and thought. Languages “cannot,” therefore, “be described as pure sign systems. By the aim usually attributed to them they are first and foremost sign systems; but by their internal structure they are first and foremost something different, namely systems of figuræ that can be used to construct signs.” For Hjelmslev, the “definition of a language as a sign system has thus shown itself, on closer analysis, to be unsatisfactory” (47).
Like Saussure, Hjelmslev sees the sign at work between the “deux masses amorphes” of “des idées” and “des sons” (Saussure, cited in Hjelmslev 49-50). Unlike Saussure, Hjelmslev argues that
there is no basis for the assumption that content-substance (thought) or expression-substance (sound-chain) precede language in time or hierarchical order, or vice versa. If we maintain Saussure’s terminology—and precisely from his assumptions—it becomes clear that the substance depends on the form to such a degree that it lives exclusively by its favor and can in no sense be said to have independent existence. (50)
And yet, Hjelmslev’s next move is to postulate a “common factor” that is the “structural principle” of all languages: the “purport,” the “thought itself” (50). The purport is an “amorphous mass,” an “unalyzed entity,” that is “ordered, articulated, formed in different ways in different languages” (50-51). Along this line of reasoning, then, Hjelmslev contends that each “language lays down its own boundaries within the amorphous “thought-mass” and stresses different factors in it in different arrangements, puts the centers of gravity in different places and gives them different emphases. It is like one and the same handful of sand that is formed in quite different patterns” (52).
There is thus “in the linguistic content, in its process, a specific form, the content-form, which is independent of, and stands in arbitrary relation to, the purport, and forms it into a content-substance,” while in the “system of expression” the “phonetic zones of purport” are “ordered to their expression-form as expression-substance” (55-56). Hjelmslev’s definition of the “sign” follows:
By virtue of the sign function and only by virtue of it, exist its two functives, which can now be precisely designated as the content-form and the expression-form. And by virtue of the content-form and the expression-form, and only by virtue of them, exist respectively the content-substance and the expression-substance, which appear by the form’s being projected on the purport, just as an open net casts its shadow down on an undivided surface. (57)
With this reformulation of Saussure’s doctrine of the sign, linguistic theory enters a new domain. It is no longer chiefly concerned with signs as entities, but with the “expression plane” and the “content plane” of language. This new focus, Hjelmslev contends, “casts light on the whole mechanism of language in a fashion hitherto unknown” (59). While in Saussure the “mechanism of language” is a “dual system,” consisting of a “reciprocating” movement between two differential functions, the syntagmatic and paradigmatic, in Hjelmslev the “mechanism of language” is defined by the “mutually opposed functives,” the content-form and the expression-form, “of one and the same function,” the sign function (Saussure 128-131; Hjelmslev 60). But is this really a new domain of linguistic theory, or just an alternative perspective on the matter? It is difficult to see how Hjelsmlev’s system actually differs from Saussure’s. Indeed, by presuming to some greater clarity, Hjelmslev only muddies the waters of the semiology, coming dangerously close to a regression into substantialism.
Having thus opened linguistics to the study of figuræ rather than signs, Hjelmslev thus moves from semiology to “glossematics” (80). He proceeds to analyze the figuræ of language into “variants” and “invariants,” proposing complementary principles of “commutation,” “permutation,” and “substitution” (62, 74). He sets out the distinction between “linguistic schema” (the “linguistic hierarchy”) and “linguistic usage” (the “non-linguistic hierarchy”) (81). He divides variants into “free” and “bound” variants—“variations” and “varieties” respectively (82). Variations that “cannot be further articulated” are “individual,” and varieties that “cannot be further articulated” are “localized” (83). He returns briefly to the language of functions to define a “sum” as a “class that has function to one or more other classes within the same rank.” A “syntagmatic sum” is thus a “unit,” and a “paradigmatic sum” a “category” (84-85). He then asserts that “functions always are present either between sums or between functions; in other words, every entity is a sum . . . each entity may be considered as a sum, namely, in every case, as a sum of variants . . . In the theory this means that an entity is nothing else than two or more entities with mutual function” (85). He quickly touches upon the “functival categor[ies]” of analysis derived from the registration “by articulation of a functional category,” such as solidarity, selection, or combination, “according to [the] functival possibilities” of the category (86). Next, Hjelmslev discusses the “phenomen[a]” of “syncretism” (87) and “catalysis” (94), the ““algebraic” entities” of linguistic science (97), “taxemes” and “glossemes” (99, 100), the “line between language and non-language” (103), the “formal definition of a semiotic” (106), and the varieties of semiotics: “denotative,” “connotative,” “metasemiotic,” “semiology,” and “metasemiology” (114). And with “metasemiology,” Hjelmslev arrives at a peculiar reversal: “metasemiology is in practice identical with the so-called description of substance” (124).
Hjelmslev presumes to have “elicit[ed] from language itself its secret” (127). Is he justified in his claim? He asserts that
Linguistic theory here takes up in an undreamed-of way and in undreamed-of measure the duties that it imposed on itself . . . In its point of departure linguistic theory was established as immanent, with constancy, system, and internal function as its sole aims, to the apparent cost of fluctuation and nuance, life and concrete physical and phenomenological reality. A temporary restriction of the field of vision was the price that had to be paid to elicit from language itself its secret. But precisely through that immanent point of view and by virtue of it, language itself returns the price that it demanded. In a higher sense than in linguistics till now, language has again become a key-position in knowledge. Instead of hindering transcendence, immanence has given it a new and better basis; immanence and transcendence are joined in a higher unity on the basis of immanence. Linguistic theory is led by an inner necessity to recognize not merely the linguistic system, in its schema and in its usage, in its totality and in its individuality, but also man and human society behind language, and all man’s sphere of knowledge through language. At that point linguistic theory has reached its prescribed goal: humanitas et universitas. (127)
This conclusion is mere posturing. What began as a prolegomena to a theory has laid claim to a “key-position” in “knowledge,” and to “man” and “human society” (127). How? What “secret” merits such grandiosity? That the sign system is actually the relation of expression and content by the sign-function? Indeed, this is not an advance upon Saussure’s thought, but a rephrasing of it. In placing the sign-function in a secondary position to the expression-plane and content-plane which it relates, Hjelmslev actually loses sight of the Saussurean innovation with which he began his text. Certainly, his claim that expression and content, signifier and signified, are each possessed of both form and substance, is intriguing, but whatever benefit such an idea might provide is obscured by his emphasis on the planes themselves, rather than the function of the relation between them. The problem of the sign remains: how does an arbitrary entity come to refer to something? Hjelmslev stumbles into challenges that exceed him. Though of historical import, Hjelmslev’s Prolegomena proves inconsistent, inattentive to its own unanalyzed premises, and unaware of its own capacities.
Hjelmslev, Louis. Prolegomena to a Theory of Language. 1943. Translated by Francis J. Whitfield, University of Wisconsin Press, 1961.
Saussure, Ferdinand de. Course in General Linguistics. 1916. Translated by Wade Baskin, edited by Perry Meisel and Haun Saussy, Columbia University Press, 2011.
Stein, Eric. “Relation over Substance: Saussure’s Revolution in Linguistics.” English 607. 2017.