“Edmund Husserl’s Phenomenology”
III Husserl’s Conception of
“The Grammatical” and
Contemporary Linguistics
Since the Middle Ages philosophers have periodically made proposals for a universal apriori grammar, frequently suggesting that such a grammar be considered as a branch or an application of formal logic. These researches have never progressed very far, not even during the period when grammarians were themselves primarily logicians. In the modern period, since scientific linguistics has vindicated its own independence of logic and philosophy, philosophical proposals of this kind have fallen into “scientific” disrepute. Thus, Edmund Husserl’s project for a “pure logical grammar”—which is probably the most recent full-scale proposal in this area from the side of philosophy—has fallen upon deaf ears. But now, within the past two decades, Noam Chomsky has begun to propose, from the side of linguistics itself, a program for the study of grammar which, if it were to succeed, might seem to justify the earlier intuitions of rationalistic philosophy and to give a new grounding to its ancient quest. Might it not be, after all, that what was needed was a more sophisticated development of grammatical studies themselves before such a proposal could be sufficiently clarified to be prosecuted with any confidence?
However that may be, we are concerned here, first of all, with Husserl, who not only neglects scientific linguistics but even the philosophical tradition. It is true that he mentions Von Humboldt1 and Scotus (Thomas of Erfurt)2 and refers in a general way to the seventeenth-century French grammarians, but he attempts to restate the problem completely independently of tradition, starting once again from the beginning, de novo. He is interested in establishing the basis for an eidetic of language within his general phenomenological reflection on our experience of speaking a language.3 Such an approach will not provide a complete or a total account of language; in fact it will be limited to an examination of certain apriori characteristics of language which might, at first glance, seem to be no more than the enumeration of a series of trivialities4 which—in their abstract generality—may seem to emasculate the phenomenon of language by reducing its enormous and known complexities and rich resources for expression to some unreal and emaciated “essence”.5
But if it should be the case that natural languages obey certain apriori laws and manifest an “ideal framework” which is “absolutely stable,” in spite of the empirical and accidental differences proper to each particular language, then the neglect of this aspect of linguistic reality would render the linguist ultimately unable to account rationally for his science. And Husserl firmly believes this to be the case:
Language has not only physiological, psychological and cultural-historical, but also apriori foundations. These last concern the essential meaning-forms and the apriori laws of their combinations and modifications, and no language is thinkable which would not be essentially determined by this apriori. Every linguist, whether or not he is clearly aware of the fact, operates with concepts coming from this domain.6
The very fact that one can meaningfully ask such questions as: How do German, Latin, Chinese, etc., express “the” plural, “the” categorical proposition, “the” hypothetical premise, “the” modes of possibility and probability, “the” negative, etc., shows the conceptual validity of such an inquiry into the aprioris of grammar.7 Against his contemporaries, among whom the sense of the apriori had “threatened, almost, to atrophy,” Husserl asks philosophers to “learn by heart” that, wherever philosophical interests are involved, “it is of the greatest importance sharply to separate the apriori.”8 We must not ignore “the great intuition of Kant.” It does not become philosophers, who are almost the sole guardians of “pure theory” among us, to let themselves be guided merely by questions of practical and empirical utility and, in the case of grammar, to allow this study to be simply parceled out among a number of ill-defined empirical sciences, since it is also governed by a framework of unified apriori laws which define its true “scientific” boundaries.
Husserl’s study of grammar locates this discipline as the first or lowest level of formal logic and states that a phenomenological approach to logic must “be guided” by language. He means by this, not that the empirical, psychological, physiological, historical, and cultural bases of language be incorporated into philosophy nor that logic is dependent on any given natural language, but rather that the study of “the grammatical” (not a given, empirical “grammar”) is the first level of logical reflection. The two primordial types of intentional experience, according to Husserl, are (1) the experience of the world and (2) the experience of language. The theoretical elaboration of the first is logically posterior to the theoretical investigation of the second, namely, language.9
To consider language in itself is to operate an implicit phenomenological reduction, i.e., to turn from the Lebenswelt of factual experience in which meanings are instantiated in factual situations to the separated meanings themselves, as they are experienced in their ideality, independently of any possible factual reference.10 The experience of language is the experience of meaning par excellence; it is our route of access to the realm of “the meant,” of “sense” and “signification.” If one distinguishes the realm of significations (what Husserl calls “categories of signification” as opposed to the “categories of the object”) from the realm of objects signified through language, one isolates within formal logic the territory of “apophantic analytics” or the purely formal study of the structures of judgment.11
Now, the first level of the implicit phenomenological reduction (if we can call it that) which is operated by the “linguistic turn” away from the world toward language itself is that of the discovery and analysis of the grammatical. Husserl calls this the study of the “pure morphology of significations” (reine Formenlehre der Bedeutungen) or “pure apriori (logical) grammar.” Such a study is strictly apriori and purely logical, a study of “the grammatical” as opposed to the empirical and historical investigation of comparative grammars; it constitutes the first level of “apophantic analytics,” to be followed by the second (the logic of noncontradiction) and the third (the logic of truth) levels of the formal analysis of signification.
No philosopher can escape the apriori rules which prescribe the conditions under which a linguistic utterance can have a unified, intelligible sense. The study of grammar, in this sense, is necessarily philosophical. Pure logical grammar (or apophantic morphology) is, according to Husserl, that first branch of formal logic which establishes the formal grammatical rules necessary for any statement to be meaningful at all; it is prior to and independent of all questions of the formal validity and the truth value of statements. Every judgment must, for instance, respect the apriori grammatical rule that in a well-formed sentence a substantive must take the place of S (in the “primitive form” S is p) and a predicate must be substituted for p. If this rule is violated, nonsense (Unsinn) results. We get strings of words like “King but where seems and,” “This frivolous green, red world,” “A man whereas our fathers,” etc., which are devoid of any unified meaning and do not assert anything as sentences or concatenations of words; the words individually may have meaning, but, when they are arranged in this way, they have none. It is the purpose of pure logical grammar to derive from the originary form of judgment (S is p) the laws which govern the formation of potentially meaningful affirmative, negative, universal, particular, hypothetical, causal, conjunctive, disjunctive, etc., forms. It is in this sense that das Grammatische selbst founds the second and third levels of formal logic and establishes rules which are always already taken for granted in the logic of noncontradiction and truth. These purely formal grammatical laws are wholly independent of the truth or falsity of the statements they rule, and guarantee only that the statements formed in accord with them will be free of Unsinn (nonsense). They have no relevance to the material contradiction (Widersinn) involved in such well-formed sentences as “Some squares are round” or “This algebraic number is green,” etc. The laws of logical grammar save us from formal nonsense only; it is the other levels of logic which save us from contradiction and countersense. However, thus to vindicate the value of pure grammatical aprioris is not to assert that logic is based on ordinary language or empirical linguistics. Husserl insists on this: Logic is grounded not on grammar but on “the grammatical.”
It is . . . not without reason that people often say that formal logic has let itself be guided by grammar. In the case of the theory of forms, however, this is not a reproach but a necessity—provided that, for guided by grammar (a word intended to bring to mind de facto historical languages and their grammatical description), guidance by the grammatical itself be substituted.12
This is grammar raised to the level of the analysis of the formal conditions of thought. It is here that Husserl joins the seventeenth-century proponents of a grammaire générale et raisonée in conscious opposition to the accepted views of his historicist and psychologistic contemporaries. The task of logical grammar is to study and furnish the apriori rules which govern the structural coherence of “parts of speech” with one another in sentences. Such grammatical rules are not just historical accidents or conventions but necessary conditions of meaningfulness and for the avoidance of nonsense; they are not, without the higher levels of formal logic built upon them, sufficient conditions for the avoidance of contradiction and error:
Nothing else has so greatly confused discussion of the question of the correct relationship between logic and grammar as the continual confounding of the two logical spheres that we have distinguished sharply as the lower and the upper and have characterized by means of their negative counterparts: the sphere of nonsense and the sphere of countersense.13
Husserl thus vindicates the place of pure grammar (reinlogische Grammatik), as a theory in its own right, within his phenomenological hierarchy of “sciences.” But it is, so to speak, the emptiest and the most formal of all. Its rules provide the barest minimal conditions necessary to avoid nonsense in forming linguistic statements. They exclude only the purest nonsense, which it would never occur to anyone to utter. Pure grammar establishes rules which are always subunderstood and already taken for granted in all the formal systems which study and establish the sufficient conditions for meaningful expressions. But the fact that the uncovering of these conditions has no “practical” value and even seems to make a science of what is trivially obvious is no reason to despise it. Its theoretical value for philosophy, Husserl tells us, is “all the greater.” Husserl takes pride in this discovery; he even glories in the fact that only philosophers are concerned with the apriori, with the discovery of truths so fundamental that all the other sciences take them for granted. It is, he believes, precisely such “obvious” trivialities as those expressed by the rules of pure grammar that mask the deepest philosophical problems, and he sees that in a profound, if paradoxical, sense, philosophy is the science of trivialities.14 The clear distinction which he was able to establish between pure grammar and the “higher” levels of formal analytics seemed to him to be a theoretical discovery of the first magnitude and a necessary point of departure for the elaboration of a phenomenological theory of consciousness.
We can best give a general outline of what Husserl means by pure logical grammar by taking his earliest discussion of this problem in the Fourth of the Logical Investigations together with his more developed discussions of the Formal and Transcendental Logic. These discussions, in turn, are but one application of the “logic of wholes and parts” of the Third Investigation. A grammatical unit is, indeed, one of the best illustrations of Husserl’s doctrine of wholes and parts:
If we inquire into the reasons why certain combinations are permitted and certain others prohibited in our language, we shall be, in a very great measure, referred to accidental linguistic habits and, in general, to facts of linguistic development that are different with different linguistic communities. But, in another part, we meet with the essential distinction between independent and dependent meanings, as also with the apriori laws—essentially connected with that distinction—of combinations of meanings and of meaning modifications: laws that must more or less clearly exhibit themselves in the theory of grammatical forms and in a corresponding class of grammatical incompatibilities in every developed language.15
A linguistic expression, whether dependent (like a word which functions as a syntactical category within a sentence) or independent (a sentence or proposition), is a string of sounds whose unity is founded in its “meaning.” Any string of sounds devoid of a unified sense (or meaning) is just that: a string of noises. What makes a string of sounds a linguistic expression is its unified meaning. A “nominal” (substantival) or an “adjectival” (predicate) expression is an example of dependent meanings; only a fully propositional meaning, which joins such dependent parts into a unified whole, is independent. The first task of logical grammar is to establish the “pure categories of meaning” as they can be related in this dependent-independent relationship.16 (It is not necessary here to follow Husserl into his detailed discussions of simple and compound meanings of these to simple and compound objects of “reference,” though this would be necessary in any complete account of his thought.)
What is important is that any linguistic expression is a “whole” composed of “parts” which are members (or “moments”) of the constituted whole rather than merely pieces (discrete elements) only incidentally and de facto attached to one another; a member of a whole obeys laws which are distinctive of the role it plays within this unified system and which are not the same as it would exercise were it, per impossibile, detached from the whole of which it is an integral part. The members of a whole interpenetrate and codetermine one another and, as such, are inseparable from one another and from the whole of which they are parts. Mere “pieces” on the contrary, would be just what they are even if separated from the whole of which they are, by analysis, found to be parts. An example of parts which codetermine one another as members of a whole would be the “extension,” “surface,” “color,” and “brightness” of a physical, perceptual object. One is not present without the other; there cannot be brightness without color, or color without surface, or surface without extension. As did Plato before him,17 Husserl considers color and surface to be related according to an apriori law (which is “synthetic” or “material” rather than “analytical” since the idea of color is not analytically contained in the idea of surface or extension) given in perception.18 Such an apriori law is not the result of my personal or cultural conditioning; it is not an empirical psychological fact about my experience, nor is it based on some statistical probability. It is a law founded in the very meaning of color and extension; what I mean by color and what I mean by extension requires that every instance of one be an instance of the other; and, once I understand this, every experience which illustrates the one will illustrate the other, and I can know this without any appeal to future experience. That brightness entails color, color entails surface, surface entails extension is an apriori law of the constitution of perceptual objects, and no act of perception will or can contradict such a law because it is part of what is meant by a physical perceptual object. Another way of stating this is to say that a physical perceptual object is a “whole” which consists of parts which are integrated into the whole as constituent members of this unified object.19
We can apply this notion to grammar immediately by noting that an independent meaning, namely, a proposition, is a formal structural whole consisting of at least a minimal number of constituent parts related to one another by apriori laws which govern their meaning-functions within the one unified whole, which is a complete, meaningful sentence. In other words, what one means by a complete, unified, independent linguistic expression (S is p) is that its parts be related to one another by apriori laws of composition which we call “syntax.”20 Dependent terms also have a unified kernel of meaning, but this meaning requires that it be completed according to certain rules if it is to function within the meaningful complex which is a whole sentence. In short, a sentence will be grammatically well-formed, and hence potentially meaningful, if and only if certain apriori rules for the correct integration of partial meanings into a whole meaning are observed. These rules are the laws of pure logical grammar, they are laws which govern the potential meaningfulness of sentences and are independent of and prior to the laws which govern internal consistency and possible truth. Meaningfulness is a prior condition for noncontradiction. The string “King but or blue” is meaningless (unsinnig), whereas the string “There are some squares which are round” is inconsistent or contradictory (widersinnig); the former, but not the latter, violates apriori and purely formal “grammatical” rules. The rules of grammar are sufficient only to guarantee grammatical coherence; they are not sufficient, though they are necessary, to guarantee logical consistency in the full sense. Pure logical grammar classifies meaning-forms and is concerned with “the mere possibility of judgments as judgments, without inquiry whether they are true or false, or contradictory.”21 Truth and falsity, according to Husserl, pertain not to propositions as such but to the laws of the assertion of propositions and thus belong to a higher level of logic.
The second step in the elaboration of a pure logical grammar (after establishing the “pure” or formal categories of meaning such as S, p, S is p, etc.) concerns the laws of the composition of partial meanings into well-formed wholes or sentences.22 At the limit, no word can be taken and defined without relation to its possible grammatical functions within a complete, unified meaning-whole; the grammatical distinctions (“parts of speech,” etc.) given in dictionaries bear testimony to this fact. Wherever there is found some grammatical distinction (or “marker”) attached to a word, this is the mark of a certain incompleteness of meaning: and thus grammatical distinctions are guides to essential meaning-distinctions within sentences.23 Sentences, unlike words, have no such “markers.”
We begin with the analysis of the pure syntactical categories. When words are combined to form sentences, they are necessarily given a syntactical “form” which permits their integration as partial or dependent meanings into a complete or independent expression. This requires that there be a restricted number of primitive connecting forms, such as the predicative, attributive, conjunctive, disjunctive, hypothetical, etc., and that there be pure syntactical forms, such as the substantive, the predicative, the propositional, etc. This is the basis for the fundamental distinction between syntactical forms and syntactical stuffs and for the recognition that the propositional form presupposes the subject form and the predicate form. Whether I take a given word as the “subject” of a sentence (“This paper is white”) or as the “object” (“I am writing on this paper”), the word—as a “term”—bears the core of meaning (and reference) which remains identical though its syntactical form varies in each case. The specific meaning and referentiality of the proposition (to a “state of affairs”) is mediated through the meaning and referentiality of its constituent terms. The proposition is a higher categorial unity “founded” on the meaning of its constituents through its giving them the syntactical form necessary to produce a unified and complete sense. Now, it can be readily seen that the number of syntactical stuffs can be infinite while the number of syntactical forms is limited and capable of complete formal definition.
Husserl calls a given unity of syntactical stuff and form the syntagma. All the members (i.e., constituent parts) of a proposition are syntagmas (and we can here neglect the analysis of the infrasyntagmatic elements or “pieces” of words and sentences, which belongs to phonology), and the proposition as a whole is also a syntagma of a higher order (i.e., “self-sufficient predicational whole . . . , a unity of syntactical stuff in a syntactical form”).24 Different members of a proposition can have the same form and different stuffs and, conversely, can have different forms but the same stuff; and these forms can be fitted into hierarchies in which what is syntactically formed on one level becomes the “stuff” of a higher form, e.g., when the proposition itself (S is p) is formed, and when it is modified modally (Is S p?, S may be p, If S is p, Then S is p, S must be p, etc.), these more complex modal forms are constructed on the basis of the Urform (S is p), which is itself composed of the infrapropositional syntagmas S (substantive), is (copular unity-form), and p (predicate).25
We must note, of course, that when a word actually occurs in a sentence, it has already been modified according to its proper syntactical form (and this form retrospectively dictates the manner in which it is defined in dictionaries under its proper “parts of speech”), since a “pure nonsyntactical stuff” is only a limit concept which can nowhere be found in actual, meaningful language (all of which is always already syntactically formed):
The forming, of course, is not an activity that was, or could have been, executed on stuffs given in advance. That would presuppose the countersense that one could have stuffs beforehand—as though they were concrete objects, instead of being abstract moments in significations.26
All the members of a proposition are “non-self-sufficient” under all circumstances; they are only what they are in the whole. I can reach the “pure stuff” of an expression only by an ideal analysis. For instance, if I examine the “syntactical stuffs” given in the sentence “This tree is green,” I am left with such words as “tree” and “green,” etc. These can be considered as “unformed stuffs,” but they are not completely unformed and thus should be called “non-syntactical forms.” For instance, if I freely vary in imagination words like “green,” “greenness,” or “similar,” “similarity,” etc., as they can appear in different syntactical forms, I reach a kernel of nonsyntactical meaning (Kernform) which remains essentially the same in its various syntactical formations; such a meaning-form “animates” a pure Kernstoff which is essentially prelinguistic—the very stuff of prepredicative experience itself. Nonsyntactical stuff (Kernstoff) and nonsyntactical form (Kernform) constitute the syntactical stuff which is “formed” by the pure laws of syntax (Kerngebilde). Nonsyntactical matter and form are only abstractions from experience; they are nonindependent constituent parts of the lowest meaningful unit, namely, the “syntactical matter” of a sentence, or what we might call the “word.”27
Now it is clear that “syntactical form” is something much more general and “formal” than syntactical stuff. If we vary different material terms like “paper,” “man,” “humanity,” “sincerity,” etc., we find that, in spite of their differences in meaning and referentiality, they possess in common an identical “form,” namely, that of “the substantive.” The same is true of “the adjectival” (which Husserl divides into “attributes” and “properties”)28 and the other basic syntactical forms.
The third and final task of pure logical grammar is, then, the construction of a closed system of basic syntactical forms and a “minimum number of independent elementary laws” for their combinations. Husserl here introduces the notion of grammatical “operation” according to which sentences can be generated. There are two interrelated tasks here which can be distinguished.
1. The fundamental forms of judgments establish laws according to which subordinate forms can be generated by derivation from the most fundamental (and, in this case, most general and abstract) forms. This is possible because the most general forms dominate the whole of pure logical grammar: The formation of a given sentence is an “operation” according to an abstract and formal rule which carries with it the law of its possible reiteration.
This moreover, should be emphasized expressly: Every operative fashioning of one form out of others has its law; and this law, in the case of operations proper, is of such a nature that the generated form can itself be submitted to a repetition of the same operation. Every law of operation thus bears within itself a law of reiteration. Conformity to this law of reiterable operation extends throughout the “whole province of judgments, and makes it possible to construct reiteratively (by means of fundamental forms and fundamental operations, which can be laid down) the infinity of possible forms of judgments.29
Thus the form S is p is more original than the form Sp is q, which is an operational transformation of it by the “operation” of converting a predicate into an attribute. These are operations, which Husserl calls “nominalization,”30 by which predicates can be transformed into substantives and also by which whole sentences can become substantives in later, derived judgments; these manifest a hierarchy of possible derivations. “This paper is white” (S is p)—“This white paper is before me” (Sp is q)—“This white paper before me is wrinkled” ((Sp)q is r), etc.31
2. The second manner in which the “primitive” form of judgment (S is p) can be transformed is through modal operations upon it. The form S is p is originary with respect to its further “doxic” modifications of the type If S is p, So S is p, Because S is p, S may be p, Let S be p, etc. Through the process of modalization the fundamental structure of the judgment (doxische Ursetzung)32 is not essentially changed; it is merely modified by special “doxic” qualities (the hypothetical, the optative, the causal, etc.), and this holds also of the more complex forms derived from the Urform (thus, If Sp is q is a modalization of Sp is q, etc.).33 It is the task of pure logical grammar to discover the basic, minimal number of laws of the derivations and modalizations of the primitive apophansis (S is p) which will account for all the possible forms of judgment which can make sense. Grammar thus is lifted up to the level of the philosophical study of language in general and becomes a part of the formal conditions of thought:
It gives us the primary and ideal structure of the expression of human thought in general, the ideal type of human language. This ideal structure (ideales Gerüst) is an exemplar or an apriori norm which defines the proper sphere of “the grammatical” (das Grammatische selbst), that is, the formal law of expressions having meaning.34
TRANSFORMATIONAL GENERATIVE GRAMMAR
Husserl thus limits himself to giving a kind of outline of what a pure logical grammar would be if it were to be worked out within his general phenomenological architectonic of interrelated and properly subordinated “sciences.” But this is sufficient to relate his project to the contemporary aprioristic approach to grammar adopted by Chomsky and his followers. We cannot recapitulate the whole theory of transformational generative grammar here, but we can perhaps outline its most fundamental presuppositions.
For Husserl apophantic morphology (pure logical grammar) is the science (or “theory”) which delimits (i.e., describes and defines) the whole infinite set of possible well-formed sentences thanks to a finite system of apriori laws which state the necessary (but not always sufficient) conditions of meaningfulness. Stated in this general way, there is an obvious similarity between what Husserl claimed could be done in the analysis of grammar and what Chomsky is in fact trying to do. They both believe that the study of grammar will illustrate certain basic laws of thought and that the “universals of grammar” are not merely the result of empirical coincidence or statistical regularities based on cross-cultural borrowing, linguistic analogies, etc., but ideal necessities of all human thought as such. For his part Husserl explicitly recognizes that there may very well be strictly empirical universals, in grammar as elsewhere, which are due to universal traits of human nature, to the contingent, historical life of the race, and that there is much in particular grammars which depends on the history of a people and even on an individual’s life-experience, but the apriori aspect of grammar (the “ideal form” of language) is independent of such empirical facts about men and culture.35 There is a slight, and perhaps, important, difference from Chomsky here, inasmuch as Chomsky wants to account not only for “formal” but also for what he calls “substantive” universals, whereas Husserl does not expect to build up a universal grammar in all its breadth but only a pure grammar which can serve as the basis for logic. Thus he admits that his apophantic morphology does not contain the totality of all the aprioris which would be relevant to universal grammar.36 In short, Husserl does not discuss the possible “material” aprioris which might be found in a phenomenological study of language; he leaves this door open.
There are more important and fundamental differences. Like Plato and Descartes, Chomsky seems to feel that precisely because it is possible to locate and describe certain apriori (and therefore universal) features of language, these aprioris must be treated as “innate” ideas or even as “biological” constituents of the human organism, a part of the DNA code. Husserl would certainly never draw such a conclusion, because it would involve him in the kind of “psychologism” he spent the first half of his philosophical life learning to avoid. Chomsky, on the other hand, is unafraid of psychologism and mentalism and freely illustrates his work (as does, for instance, Merleau-Ponty from a different perspective) with what is known about the psychological processes involved in the acquisition of language; he concludes that these facts point toward the existence in the human mind (and therefore in the human species) of a categorial structure (“linguistic competence,” innere Sprachform) which would be unlearned, innate, and temporally as well as logically prior to experience. But if his notion of “linguistic competence” can be divorced from the Cartesian theory of “innate ideas,” as I think it can be (though I cannot argue all this here), we need not tarry over this difference from Husserl.37 If it is possible, in short, to interpret the “formal universals” of language which constitute the base rules of deep grammar as aprioris in Husserl’s sense, then we can easily separate the essence of Chomsky’s work from the Cartesian folklore in which it is imbedded in his own writing.
In fact the notion of “linguistic competence” Chomsky is attempting to elaborate is based on the very straightforward linguistic fact that native speakers and hearers of a language can produce and recognize on the proper occasions an infinitely varied number of appropriate and new sentences for which their empirical linguistic habits and experience up to any given point can have prepared them only in the most abstract and schematic manner. Moreover, most speakers of a language, i.e., those who know how to speak grammatically and how to distinguish grammatical from ungrammatical sentences in their language, are not aware on the level of conscious reflection just which grammatical rules enable them to give definite interpretations to ambiguous sentences, nor can they in general explicitly state the rules which enable them to formulate and distinguish well-formed from deviant utterances. Most speakers thus operate according to a complex system of hierarchically ordered linguistic rules (which must be applied in series) without explicit awareness of just what these rules are; these rules must therefore be a subconscious possession of the speaker of a language (and in fact we know that the grammar of no natural language has been completely and explicitly codified up to now.)
Chomsky’s great originality has been the elaboration of a theory about the deep structure common to all languages and the transformational rules by which this deep structure can be converted into the phonological surface structure of given, natural languages. “The central idea of transformational grammar,” he writes, “is not only that the surface structure of a language is distinct from deep structure but that ‘surface structure’ is determined by the repeated application of certain formal operations called grammatical transformations’ to objects of a more elementary sort.”38 In short, a given sentence can be studied either from the point of view of its physical shape as a string of sounds or morphemes or from the point of view of how it expresses a unit of thought, and the latter is not adequately accounted for by the surface arrangement and phrasing of its component parts. Sentences with very similar surface structures can be seen to require very different grammatical interpretations (as, for instance, “I persuaded John to leave” and “I expected John to leave”).39 We are not interested just here in the intricacies of Chomsky’s analyses, and philosophers will probably grant him more readily than structural linguists will that the deeper “logical form” of sentences is frequently belied by their surface grammatical forms. This kind of distinction between surface and deep grammar is exactly what Husserl was aiming at when he distinguished “the grammatical” from empirical grammars,40 though he nowhere anticipated the spectacular developments in linguistic theory Chomsky has initiated without him.
We must limit ourselves here to a brief account of the nature and structure of deep grammar as Chomsky postulates it. In order to account for the full range of the infinitely variable “new” sentences we are capable of producing and recognizing on the surface level, there must be a highly restricted and hierarchically ordered system of recursive rules (what Husserl called “reiterable operations”) which constitute, in fact, the deep structure of language and, then, a set of transformational rules which can account for the production of the surface level. The transformational rules differ for each natural language; but what Chomsky calls base rules (which establish the basic grammatical categories and subcategories and the rules of their combinations) are “formal universals” common to all languages. They are “the universal conditions that prescribe the form of human language . . . ; they provide the organizing principles that make language learning possible.”41 But to say in this way “that all languages are cut to the same pattern”42 is not necessarily “to imply that there is any point-by-point correspondence between particular languages”:
To say that formal properties of the base will provide the framework for the characterization of universal categories is to assume that much of the base is common to all languages. . . . Insofar as aspects of the base structure are not specific to a particular language, they need not be stated in the grammar of this language. Instead, they are to be stated only in general linguistic theory, as part of the definition of the notion “human language” itself.43
Thus there can be no language which violates such basic universal rules, but not all of these rules need be explicitly incorporated into every natural language; we are dealing with formal, apriori conditions only. Of course, one must ask how these rules are discovered and elaborated, and the answer can only be by reflection on some one or several known languages. Though Chomsky does not say so explicitly, the method he employs would seem to be a variant of the method Husserl called “eidetic intuition,” namely, argument on the basis of examples chosen from empirical experience: a free variation and comparison of a number of examples sufficient to give one an eidetic insight into the essential structure of what is being examined, in this case linguistic behavior. We cannot directly inspect “linguistic competence,” Chomsky admits; and the very existence of the deep structures by which “competence” is described and defined must be “theoretical.” But it is not necessary to know whether the details of Chomsky’s theory are true in order to understand what it means and how it is to be elaborated as a “working hypothesis” which, at the limit, would account for the phenomenon of language in all its generality.
Here there are parallels with Husserl’s approach which are striking. Whether we attempt to explain the “linguistic competence” of a native speaker-hearer or attempt to thematize the deep structure which is this competence, there are apparently no “inductive procedures of any known sort” which we can follow.44 Certainly a speaker’s “internalized grammar. . . goes far beyond the presented primary linguistic data and is in no sense an ‘inductive generalization’ from these data”:45
It seems plain that language acquisition is based on the child’s discovery of what from a formal point of view is a deep and abstract theory—a generative grammar of his language—many of the concepts and principles of which are only remotely related to experience by long and intricate chains of unconscious quasi-inferential steps. . . . In short, the structure of particular languages may very well be largely determined by factors over which the individual has no conscious control and concerning which society may have little choice or freedom. On the basis of the best information now available, it seems reasonable to suppose that a child cannot help constructing a particular sort of transformational grammar to account for the data presented to him any more than he can control his perception of solid objects or his attention to line and angle. Thus it may be that the general features of language structure reflect, not so much the course of one’s experience, but rather the general character of one’s capacity to acquire knowledge.46
Even if one hesitates to jump to Chomsky’s conclusion that such considerations as these necessitate the postulating of “innate ideas,” one might well be tempted to give them the weaker Husserlian framework. According to Husserl, “fact” and “essence” are inseparable in experience. Every fact, in order to be understood, must be brought under an eidetic law which defines its essential meaning-structure,47 and thus linguistic facts must exemplify essential and necessary apriori structures no less than perceptual facts. It would seem that nothing essential is lost to Chomsky’s theory if its “universals” are understood as eidetic aprioris of the kind discussed by Husserl.
There is a further point. If there are eidetic or apriori structures of language as such, it ought to be possible, at least theoretically, to establish such structures on the basis of even one well-selected example, a single instance of the apriori law in question, since no instance of the phenomenon in question could fail to illustrate its essential and necessary structure.48 And we find that Chomsky makes a claim for his theory similar to this well-known Husserlian axiom. He writes:
Study of a wide range of languages is only one of the ways to evaluate the hypothesis that some formal condition is a linguistic universal. Paradoxical as this may seem at first glance, considerations internal to a single language may provide significant support for the conclusion that some formal property should be attributed not to the theory of the particular language in question (its grammar) but rather to the general linguistic theory on which the particular grammar is based.49
And also, like Husserl, Chomsky believes that the aprioris of grammar (“the grammatical”) reveal the structures of thought itself:
The central doctrine of Cartesian linguistics is that the general features of grammatical structure are common to all language and reflect certain fundamental properties of the mind. . . . There are, then, certain language universals that set limits to the variety of human language. Such universal conditions are not learned; rather they provide the organizing principles that make language learning possible, that must exist if data is to lead to knowledge.50
If we were able to examine the claims of transformational generative grammar in greater detail than we can permit ourselves here, we would be able to bring out a number of theoretical claims which appear to be just as Husserlian as we should expect on the basis of these general methodological statements. Let us limit ourselves here to the parallel discussions of “nominalization” we find in Husserl and Chomsky. Various kinds of nominalizations are Husserl’s most frequent and sustained examples of the fundamental kinds of operations which can be applied to judgments and judgment forms. There is a whole hierarchy of such possible operations. There is, first of all, the “operational transformation . . . of converting a predicate into an attribute”51 through which what had been a predicate in a proper judgment becomes absorbed into the substantive as a determining characteristic which is no longer affirmed but simply presupposed as the basis for further predication. Furthermore, any predicate (any “adjectival”) form can be nominalized and become the subject of further judgments itself (e.g., “The quality p is appropriate to S,” “The green of this tree is beautiful,” and so on).52 Finally, and more important, the proposition itself (and through it the state of affairs to which it refers) can be nominalized and thus become the substrate for a new judgment (“The fact that S is p” becomes the subject of a further predication). This is possible because, in the most fundamental sense, the primitive form of all judging (S is p) is itself “an operation: the operation of determining a determinable substrate”53 and, as such, can always be reiterated and thus generate higher and more complex forms having the same (though now hidden) formal structure. In this way the primal form can generate the infinite set of possible sentences. If we attend only to the form of propositions, we leave aside whatever complexities might be discovered by a material analysis of the terms of an actual judgment and grasp the subject (S) of the judgmental operation as a “simple object,” ultimately just as “something” or “one” (Etwas überhaupt), as subject to determination in general (without specifying the particular kind or appropriateness or validity of any particular determination other than to say that any given determination, whatever it may be, must be compatible with the sense of the subject).54 It is on the basis of these considerations that Husserl affirms within pure logical grammar the “pre-eminence of the substantival category.”55 Adjectives (predicates, whether relations or attributes) can always be substantivized, Husserl shows, whereas the converse is not the case. The proposition, as the operation of determining a determinable substrate, is necessarily ordered in terms of its substantival member, and an analysis of the manner in which a predicate can be chosen as a determination of a given subject form must be established on each level of apophantic analytics, i.e., on the level of analytical noncontradiction, and on the level of possible referential truth. The “pre-eminence of the substantive” thus expresses an absolutely fundamental structure of the logic of discourse.56
Now, if we turn to the claims of transformational grammar, what do we find? Chomsky believes that the Port-Royal grammarians were the first to discover the distinction between deep structure and surface structure as well as some transformational rules for converting semantically significant structures of the base (deep structure) into the more derived surface structures in which their true, underlying form is obscured. In their analysis of the derivations of relative clauses and noun phrases which contain attributive adjectives, the Port-Royal grammarians postulate a recursive (“operational”) rule in the base such that each relative clause and each modified noun phrase is derived from a propositional structure which is essentially the same as Husserl’s most abstract form (S is p). “The invisible God created the visible world” is, on the surface level, an implicit way of saying that “God, who is invisible, created the world, which is visible”; and this structure in turn implies a series of propositions such as: “God is invisible,” “The world is visible,” “God created the world,” and so on. The most abstract underlying structure (S is p) is what determines the semantic interpretation of the surface structure, and each relative clause and each modified noun phrase (which is but a further derivation in the same line) has a proposition at its base:
The principal form of thought. . . is the judgment, in which something is affirmed of something else. Its linguistic expression is the proposition, the two terms of which are the “sujet . . .” and the “attribut . . .” In the case of. . . the sentences just discussed, the deep structure consists of a system of propositions. . . . To form an actual sentence from such an underlying system of elementary propositions, we apply certain rules (in modern terms, grammatical transformations). . . . It is the deep structure underlying the actual utterance, a structure that is purely mental, that conveys the semantic content of the sentence. This deep structure is, nevertheless, related to actual sentences in that each of its component abstract propositions . . . could be directly realized as a simple propositional judgment.57
Thus we see that what Husserl discussed in terms of “nominalizations” receives an interpretation in terms of the transformational rules which derive surface structures from the more universal structures of the base. But transformational theory makes at least one more claim which goes a bit beyond this one. Chomsky argues that, in deep structure, noun phrases which are subjects logically precede verb phrases and that verb phrases are subject to selectional rules determined by the nouns. Though verbs can be “nominalized,” nominalization is a transformational process of mapping deep structure onto surface structure and does not affect the essential and necessary distinction between the class of nouns and the class of verbs. Like Husserl, Chomsky requires that every sentence have a subject and a predicate and that the former logically determine the selection of the latter. Nouns thus enjoy logical priority over verbs; and no language, it is asserted, is thinkable which would not contain nouns and which would not give nouns logical priority over verbs in such wise that one cannot select verbs prior to the selection of the nouns which they must modify. There is thus some kind of ontological structure to language, in its possible relation to its own referential use, which parallels the necessary perception of the world in terms of “objects.” Moreover, there are strict context-free subcategorization rules operative on the selection of nouns themselves. These are rules of the base and therefore have some universal validity and coerciveness, according to transformational theory. A noun can be either a count or a noncount noun; only if it is a count noun can it be animate or inanimate; only if it is an animate noun can it be human or nonhuman; only if it is not a count noun can it be abstract, etc. :
There is a binary choice at each stage, and the derivation is hierarchial because the rules impose an ordered set of restrictions on the syntactic features which can be associated with the nouns and limit the classification of nouns to the possibilities enumerable by the rules. As a decision in linguistic research this implies that the optimum representation of the grammar of any language contains these rules. They identify an aspect of the mechanism of language use which is fundamental.58
Thus we see that transformational grammar has discovered a way of making explicit the kind of universal conditions on grammar which a philosophy of language affirming the “pre-eminence of the substantival” might expect.59 At the present state of linguistic research it would be more hazardous to draw the parallel any further.
In conclusion we cannot consider all the arguments which have been or might be brought against this unified conception of apriori grammar; but, granting ourselves that the unity of purpose we have discerned behind the grammatical projects of Husserl and Chomsky is acceptable, we can, perhaps, touch on one typical argument we find in the writings of Merleau-Ponty vis-à-vis Husserl60 and in the writings of “structural linguists” like Hockett vis-à-vis Chomsky.61 This argument is based on the diachronic development of language through time, an evolutionary development which subjects languages to all the vicissitudes of cultural history. The vast proliferation of historical human languages and their known diachronic changes, it is argued, renders highly dubious the claim that there is some fixed universal, nonhistorical structure of language independent of the surface structures of the natural historical languages which I speak and which I learn. It is not only that there are gradual but never ceasing changes in sound structure and phonology, but the very forms and senses of words also change; and if we compare languages over a long period of time, we can see fundamental changes in (surface) syntax as well. (English syntax is no longer what it was in the days of King Alfred or Chaucer; French and Italian do not have the syntax of Latin, etc.) Are not such fundamental and apparently all-pervasive historical changes—to which all natural languages are subject—sufficient to cause us to reject the rationalistic hypothesis of a universal logical grammar? The fact that many languages, like Chinese and Bantu, seem to lack the subject-predicate structure of the Indo-European languages studied by Husserl and Chomsky has led some linguists to argue that these languages have a “grammar and logic” different from that of the Indo-European languages and even that they escape the logical categories of Aristotle and Leibniz altogether.62
To this challenge, we can give here no more than a schematic answer but one which, if true, is sufficient to meet it. If we restrict ourselves to the fundamental Husserlian distinction between empirical grammars and “the grammatical itself,” I believe Husserl would (and Chomsky could) point to at least one structure of language which would resist the thrust of this “empiricist” observation. There are certain linguistic facts—such as the translatability of all natural languages into one another, the recognition that every natural language is sufficient for all purposes of human expression and that none is privileged, that anything which can be said in any language can, in principle, be said equally well in any other—which point in the rationalistic direction. It is not necessary that there can be no loss of meaning in the movement of translation from one language to another (clearly the levels of meaning tied directly to phonological systems and even to morphophonemics are only incompletely translatable) but only that some categorial level of identifiable sameness of meaning be reproducible in any natural language. This, Husserl would say, is primarily the unit of meaning carried by the syntactically well-formed sentence. A sentence which is formed in accord with the fundamental apriori laws of signification must necessarily have a sense, and this sense must necessarily be one. This is the apriorisches Bedeutungsgesetz which normatively determines and guarantees the possibility and the unit of a given independent meaning in his sense.63 There can be no language, he would argue, which is not formed on the basis of units of meaning (i.e., sentences as “wholes” composed of syntactically formed “members”), because this is what is meant by language. It is these units of meaning which are—in the primary sense—translatable from one language to another and, in principle, expressible in any. Certainly the manners in which the Bantu, the Chinese, the Semitic, and the Indo-European languages, for instance, express the various kinds of propositions differ as to their morphology, and there will be some languages whose morphology will explicitly incorporate forms for the expression of meaning that other languages must express in some other (perhaps nonmorphological) manner. But would we call a language a human language if it had no means of expressing the units of independent and unified meaning which can be thematized in logical propositions? If we answer this question negatively, we will have recognized the fundamental form of linguistic meaning from which all other possible forms can be derived; and this recognition will not be based on statistical probabilities or on an appeal to future experience but will be a conceptual or “eidetic” claim about the nature of language and thought as such.
We use cookies to analyze our traffic. Please decide if you are willing to accept cookies from our website. You can change this setting anytime in Privacy Settings.