Chomsky versus Jackendoff; Generative and Interpretive Semantics

Posted by:


This paper is an examination of generative and interpretive semantics which are the two controversial approaches to the study of semantics. It is observed that generative semantics as championed by Chomsky firmly stated that semantics (meaning) cannot be separated from syntax while the interpretive approach arguably, tried to separate syntax from semantics. The transformational rules and selection restrictions which are tenets of generative semantics were examined. It is also noted that despite these approaches, the meaning of meaning still remains a strong debate and neither of the approaches sufficiently captures “meaning”.
Keywords: Generative Semantics, Interpretive semantics, syntactic structures, deep structures, lexicon, surface structures.


Generative Semantics is a theory of generative grammar which sought to explain the concept of “meaning”. Chomsky’s opinion is that the deep structure of a sentence is equivalent to its semantic representation. This means that surface structure (syntax) is simply representing what is already derived in the deep structure.
According to Chomsky (1976, p.62) “A grammar of a language can be loosely described as a system of rules that express the correspondence between sound and meaning in this language”.
Chomsky views language as a system of rules. This means that there are rules guiding the generation of words and there are equally rules guiding the representations of these items which are generated in the lexicon (deep structures) and are map onto the surface through phonology and syntax. The rules guiding these transformations are called transformational rules.
Generative grammar sought to relate meaning with syntax and sound through a set of transformations from deep structure to surface structure
According to Nwala (2015, p. 206)
This approach (Generative Semantics) is championed by Chomsky with his theory of innatism. The concept of innatism has it that the meaning of a language is generated in the mind. The mind posses a definite set of rules which converts different kinds of sentences and determine the meanings in the inner state called the deep structure.
This means that meaning is derived from the deep structure which is realizable through transformational rules. Hence, Chomsky’s Aspect (1957) portrayed “deep structure” as directly connected to meaning.

The program developed out of transformational generative grammar in the mid 1960s. Chomsky is of the view that the concept of meaning is realized from “deep structures”. This position sparked a lot of debates because the concept of “meaning” has been an ongoing controversy since time immemorial. Philosophers have argued and rationalized on the concept of meaning. Some linguists say meaning is “open” while some believe that meaning is “individualistic”; still they have not been able to arrive at a definition of meaning that is acceptable to all. Hence, Chomsky’s opinion that meaning is realized from deep structure has only added to the ongoing debate.
Throughout the late 1960s and 1970s, there were heated debates between linguists who champion the views of generative semantics and those in favour of interpretive semantics.
According to Umera- Okeke (2015, p.106) “In studying semantics in generative grammar there are two basic approach (1) Generative Semantics (2) Interpretative Semantics”
Interpretive Semantics is of the view that syntactic rules enumerated a set of well formed sentences paired with syntactic structures, each of which was assigned an interpretation by the rules of a separate semantic theory. This approach (interpretive) clearly separated syntax from semantics.
According to Bagha (2011) Interpretive Semantics as “an approach to generative grammar assumes that rules of semantics interpretation apply to already generated syntactic structure”
(Cited in Umera-Okeke 2013, p.107)
Generative semantics on the other hand is of the view that interpretations were generated directly from deep structures and were subsequently transformed into recognizable sentences by transformations.
In corroboration to this stance (Nwala 2013, p.206) opined that:
Meanings of words are registered in the lexicon which is the base of a grammar and part of the deep structure, and are lifted and map unto sentences via transformational rules.
Also Chomsky (1976, p. 65) opined that:
“It is natural (though I shall argue, only in part correct) to suppose that the semantic interpretation of a sentence is determined by the intrinsic semantic content of lexical items and the manner in which they are related at the level of deep structure”.
This simply means that deep structures determine semantic representation under the rules of semantic representation.
A lot of controversy surrounds generative semantics because of its argument that syntax and semantics cannot be divorced. Scholars like Jackendoff opine that there should be a separate account of semantic rules and representations just as there are accounts of syntactic rules and accounts of phonological rules.
In arguing against those in favour of a separate semantic theory, Chomsky (1976)stated:

Clearly… there is no empirical difference between the ‘syntactically-based’ Standard theory and the ‘semantically-based’ alternative. The standard theory generates quadruples (P,S,d,S) (P – a phonetic representation, S- a surface structure, d- a deep structure, S-a semantic representation). It is meaningless to ask whether it does so by ‘first’ generating d, then mapping it unto S (on one side) and onto S and the P (on the other side); or whether it ‘first’ generates S (selecting it, however one wishes, from the universal set of semantic representations), and then maps it unto d, then S, then P; or for that matter, whether it ‘first’ selects the pair (P, d) which is then mapped unto the pair (s, S); etc. At this level of discussion, all of these alternatives are equivalent ways of talking about the same theory. There is no general notion ‘direction of a mapping’ or ‘order of steps of generation’ to which one can appeal in attempting to differentiate the ‘syntactically-based’ standard theory from the ‘semantically-based alternative (p.69-70).

From the above, it becomes obvious that Chomsky’s argument is that meaning is generated from the deep structures and then mapped unto the surface structure through phonetic representation (P) and syntactic structure (S). He is therefore of the opinion that advocating for a different semantic theory is meaningless since meaning is generated in the deep structures and syntactic representations are equally generated from the deep structures, hence, semantics cannot be separated from syntax. This proposes that his Standard Theory sufficiently accounts for meaning (semantics).
The theory of Generative semantics claims that syntax and semantics are inseparable and homogenous. It means that semantic factors enter into the formulation of transformational rules. Therefore meaning is accounted for in syntax.
Chomsky’s Standard Theory and the later Revised Extended Standard theory is based on the notion that the deep structure of a sentence and the meanings of words (lexical items) used in that structure represent the total meaning of the sentence. At the level of deep structure, lexical items are inserted into syntactic forms, with the application of “selection restrictions”, and concepts such as subject and object are defined. Selection restrictions are rules regarding the permissible combination of lexical items in language. These rules prevent the generation of un-meaningful or anomalous sentences such as “colorless green ideas sleep furiously” or combinations such as “red hope”.
(Syal and Jindal 2010, p. 154)
Under rules of selection restrictions sentences such as:
a) The pizza ate the man
b) The firewood skipped town
c) The dog spoke in parables
Cannot be formed because they simply do not conform to our knowledge of the world. This simply means that meaning is derived from the deep structure and any item we select from the lexicon we do so bearing the meaning we intend in mind and “selection restriction” helps to deter us from making ungrammatical sentences since they only allow “Permissible combination of lexical items.”
To buttress this point let’s examine further.
For example
A) Transitive Verb
A transitive verb is a verb that takes an object. This means that the action of a subject falls on the object of the sentence.
Hence in using a transitive verb, the rules of selection restrictions would restrict one from making a sentence such as:
The lady died (died is an intransitive verb)
The lady killed a goat
Subject verb object
Since the verb “killed” is a transitive verb and would take an object (usually a noun phrase) in the syntactic structure. Thus, the sentence “the lady died” can be generated but the sentence “the lady killed…” cannot be generated alone unless there is a noun phrase after it to give it a complete meaning.
B) The verbs “Scribble” and “Write” both means to “Scrawl”. They are in the same category; this means that they have the same selection restrictions.
We cannot say for instance:
– The tank scribbled some letters on the woman or
– The tank wrote some instructions on the woman.
Instead we can say:
– The woman scribbled some letters on the tank
– The woman wrote some instructions on the tank.
The last two examples are acceptable. They belong to the same semantic relation. They are synonyms. Also,
The “tank” cannot write or scribble because it is
The “woman” can perform such action because woman is
Therefore the “tank wrote” or the “tank scribbled” is not permissible. Hence the knowledge of rules regarding selection and specific properties of lexical items are innate. This means they are present in the internalized lexicon of a language which every speaker possesses. This internalized lexicon which every native speaker possesses is part of what Chomsky called “deep structures”
According to Chomsky (1987) deep structures in generative semantics theory are held to meet several conditions:
One: they determine semantic representation
Two: they are mapped into well-formed surface structures by grammatical transformation
Three: They satisfy the set of formal conditions defined by base rules; in particular, the rules of the categorical component define the grammatical functions and order of constituents, and the contextual features of lexical entries determine how lexical items can be entered into such structures.
These are observable in the instances given above.
However, some linguists do not agree with Chomsky, for them semantic meaning is not dependent on the syntactic structure. Linguists like Jackendoff took a different approach by adopting a different hypothesis – that if rules of semantic interpretation can be formulated properly, their properties and the properties of the semantic representations they derive can equally be used to account for the concept of semantics by leaving the syntactic component as free of semantic intervention. This approach is called the interpretive approach to semantics. (Jackendoff 1972)
According to Jackendoff (1972 p. 1)
“In the early days of generative grammar, the nature of the rules relating syntactic structures to meaning was not discussed. Chomsky’s Syntactic structures (1957) shows that a linguistic theory in which meaning is determined at least in part by a level of underlying Structure can capture important generalizations. But Chomsky does not propose explicit Mechanisms for representing or deriving meaning; his main concern is with the formal syntactic devices of the language”
Things however took a different turn with the publication of Katz and Fodor, The Structure of a Semantic Theory (1963). They stated that a grammar should be thought of as a system of rules linking the external or surface form of the sentences of a language to their meaning. This means that for any linguistic analysis to be complete it must account for meaning. Katz and Fodor suggest that meanings are to be expressed in a universal semantic representation just as sounds are expressed in a universal phonetic representation such as the IPA.
Unlike Chomsky’s view, some linguists agree that semantic representations are not similar to syntactic structures. Jackendoff’s approach to semantic representation was separated into four parts, including two hierarchical structures.
The first hierarchical structure was
1. The Functional structure: this represents relations in the sentence induce by the verbs, including such notions as agency, motion, and direction.
The second hierarchical structure was
2. The modal structure: This specifies the conditions under which a sentence purports to correspond to situations in the real world.
3. The table of conference: This indicates whether pairs of noun phrases in the sentence are intended to be coreferential or not.
4. The focus and presupposition designate what information in the sentence is intended to be new and what is intended to be old.
According to Jackendoff (1972) the failure of earlier studies to properly distinguish these semantic substructures, particularly the two hierarchical structures, has been the source of much difficulty and confusion (p.3)
Likewise Katz and Postal (1964) proposed a semantic component which states that the only syntactic information used in determining semantic representation is the underlying (deep) structure. They assume that functional structure is the sole source of semantic information.
Sample of Katz and Postal’s Proposal

Semantic representation
Semantic Component
Base rules Deep structures
Transformational component
Surface Structures
Jackendoff’s Alternative
Base rules Deep Structures Functional Structures
Transformational Modal Structures
Component and Table of Coreference
Surface Structures Focus and Presupposition

Jackendoff states that when all of these are put into consideration, an account of semantics which is empirical can be realized. This approach goes beyond the consideration of lexical items alone. It considers pronouns as anaphors (referents) hence the concept of “coreference” which is a concept that means a situation where two or more expressions in a text refer to the same person.
E..g. Lilian wanted to come but she changed her mind.
b) John wanted to help but the bastard didn’t show up
c) she didn’t come, she was too lazy to make it
The sentences above are coreferential. All refer to the same referent i.e. the pronouns are the same with the referents. Jackendoff advocates for interpretive semantics arguing that transformational rules have no way of accounting for coreference. (p. 109)
Jackendoff’s approach differs from Chomsky’s Generative Semantics in that Jackendoff claimed that meaning has an inventory of its own which is not dependent on the syntactic Structure. He said:
“Meaning has an inventory of basic units and of means to combine them that is as distinct from syntax as syntax from phonology”
(Jackendoff 2003, p.124).
He further stated that:
In short, we come to see semantics not as derived from syntax, but as an independent generative system correlated with syntax through an interface. It instantiates the thoughts that language expresses; syntax and phonology are the means by which thoughts are converted into overt expressions.
Here, Jackendoff simply asserts that meaning is an independent system which finds expression through syntax and phonology but meaning is not dependent on syntax.
Chomsky portrayed deep structure as directly connected to meaning. He stated that meaning (Semantics) cannot be separated from syntax. He said:
There is no general notion ‘direction of mapping’ or ‘order of steps of generation’ to which one can appeal in attempting to differentiate the ‘syntactically based’ Standard Theory from the “Semantically based” alternative view or either from the “alternative view” which regards the pairing of surface structure and semantic interpretations as determined by the ‘independently selected’ pairing of phonetic representation and deep structures etc. Before one can seek to determine whether grammar is ‘syntactically based’ or ‘semantically based’ (or whether it is based on ‘independent choice’ of paired phonetic representation and deep structure etc.) one must first demonstrate that the alternatives are genuine and not merely variant ways of speaking in a loose and informal manner about the same system of grammar. This is not so easy … (1976, p.70)
From the above statement, it becomes apparent that Chomsky’s view of meaning as inseparable from syntax is solid. He is saying that it is not easy to separate the two (semantics and syntax) and that such an endeavor is not easy to accomplish.
On trying to interpret context as part of meaning, Chomsky is saying that even context cannot be interpreted independently and that such an effort is misguided. He said:
One might argue that non linguistic beliefs, intentions of the speaker, and other factors enter into the interpretation of utterances in so intimate and perhaps so fluctuating and indefinite-a fashion that it is hopeless and misguided to attempt to represent independently the “purely grammatical” component of meaning. (Chomsky 1976, p.67)
However, scholars such as Jackendoff, Katz, Fodor and others have tried to provide alternative means to the study of meaning, still the study of meaning remains inconclusive. The debate is still on-going and scholars are yet to agree on any approach as more acceptable than the other. No side has won the debate; not the proponents of generative semantics or those championing the interpretive semantics.

This paper tried to examine the concept of generative semantics and interpretive semantics as the two approaches to the study of semantics within the tenets of transformational generative grammar (T.G.G). The interpretive school approach contrasts with the Chomskyan approach in that the former believes that there is a distinction between syntax and semantics while the later is of the opinion that semantics (meaning) is dependent on the syntactic structure. This ultimately means that a change in structure leads to a change in meaning. Chomsky concludes that any attempt to separate semantics from syntactic structure is hopeless and misguided not to mention that it is not an easy task. Meanwhile Jackendoff, Katz, Fodor came up with alternative means of studying semantics as distinct from syntax and Jackendoff concludes that meaning (semantics) is different; it only finds expression through syntax and phonology.

Chomsky, N. (1976). Studies on semantics in generative grammar. The Hague: Mouton
Fodor, J. (1975). The language of thought. Cambridge Mass: Harvard University Press.
Jackendoff, R.S (2003) Foundations of language: Brain, meaning, grammar, evolution. New York: Oxford University press.
Jackendoff, R.S (1972) Semantic Interpretation in Generative Grammar. Cambridge: M.I.T press
Katz, J. (1966). The philosophy of language. New York: Harper & Row
Nwala, M.A (2015). Introduction to Linguistics: A first course. Port Harcourt: Wisdom publishers limited.
Syal, P. & Jindal, D.V. (2010) An introduction to linguistics: Language, grammar and semantics. New Delhi: Prentice hall of India learning.
Umera-Okeke, N. (2015). Semantics: The theories of meaning in English. Awka: Fab Educational books.


About the Author:

  Related Posts
  • No related posts found.


  1. Kingsley  April 5, 2019

    It is a well written piece

    • Lilian  April 5, 2019

      I’m glad you enjoyed it


Add a Comment