Download Introduction to English Sentence Structure: Word, Phrasal, and Clausal Categories and more Lecture notes Morphology and Syntax in PDF only on Docsity!
Mohamed V University, Faculty of Letters and Human Sciences, Department of English.
Lectures on Hierarchical Constituent structure.
Elhassan Souali
- The hierarchical nature of syntactic structure In today's lecture, we are going to discuss and argue for the well- known claim in Generative Linguistics that all sentences in English (and other human languages) have a hierarchical/tree-like constituent structure, i.e., a syntactic structure consisting of hierarchically related syntactic constituents, as shown by the following example and its corresponding tree diagram (which we shall modify later, in the light of X-bar theory):
- a. These clever boys always drink milk in the morning. b. S NP VP Det AdjP N AdvP V NP PP Adj Adv N P NP Det N These clever boys always drink milk in he morning Notice that this hierarchical structure consists of three types of constituents (which is the case for all sentences in English and other languages). These are namely, word-level (or word-level) constituents, e.g. N, V, Adj, Det, ..), phrasal categories (e.g. Noun Phrase (NP), Verb Phrase (VP), ...) and clausal categories (namely the categories S, as in 1b above, and S-bar, which stands for the full clause, i.e., S plus Comp (a complementizer (or subordinating conjunction),
see below for examples and discussion). In Grammar 3, you dealt with all these constituent types from a descriptive point of view. You learned to identify them, state their grammatical form (e.g., N, NP, V, VP, ...), their grammatical function (e.g., Subject, direct object, complement, modifier, adverbial, conjunct, ...), and the sentence element with which each of these constituents is related. This means that what your previous Grammar3 course is crucial for this course of Syntax, since it is based you learned in it. What distinguishes this course, though, is that it will present to you a more up to date conception of syntactic constituents (or categories) and their internal syntactic structure, a conception which is based, as we shall show you, on both empirical and conceptual evidence. Clearly, therefore, one fundamental property of our syntax course concerns linguistic argumentation, i.e., the process whereby an assumption or claim is supported by empirical (or conceptual) evidence or arguments. For example, in today's lecture (and maybe the next one as well) we shall give linguistic evidence supporting the existence of the above- mentioned three types of syntactic categories and in the meantime providing linguistic evidence supporting the claim that sentences in English (and other languages) do have a hierarchical (not flat) syntactic structure.
- Linguistic arguments for the three types of syntactic categories/constituents 2.1. Arguments for word-level categories The first type of categories we should argue for are the so-called word-level categories (traditionally referred to as parts of speech), that is the lowest categories in syntactic structure that immediately represent words, e.g. the categories N, V, Adj, Adv, Det, ....., and which therefore regarded as the minimal/lowest units of syntactic structure, since they have no internal syntactic structure (though they have morphological and phonological structure). A better way to refer to them is by calling them X-zero categories (as opposed to XP categories; see my forthcoming lectures on X-bar theory), since this allows this better account for the fact that these lowest categories can represent not only full words but also syntactically active overt or null
The second argument for word-level categories is morphological in nature. Thus, it is well-known that the morphological properties of a word in any language (both derivational and inflectional) is largely determined by its category. Thus, the affix 'ed' in English can be attached only to lexical verbs, the plural affix 's' can attach only to nouns, the comparative affix 'er' can attach only to adjectives or adverbs, the derivational affix 'ly' is attached only to an adjective to form an adverb, etc. Like lexical categories, different functional categories have different morphological properties. For example, functional categories in English, unlike lexical categories, may be realized as overt affixes (as is the case with T with a past tense feature) or as covert ones (as is the case with T carrying the present tense feature in first and second person). And with functional categories, each type has its specific morphological properties. For example, T in English, unlike Det and Comp, can be morphologically attached only to verbs. Comp can attach only to auxiliary verbs in English, not to full verbs (in direct questions; see my forthcoming lecture on T-to-C movement). And Det is generally realized as a free morpheme in English, as opposed to languages like Arabic where it is realized as a bound morpheme). A further argument in support of the existence of different types of word-level categories, as shown above, is syntactic in nature. In other words, the syntactic behavior and distribution of a word (or syntactically active bound morpheme) is largely determined by its syntactic category). For example, Nouns syntactically function as heads of NPs, Verbs as heads of VPs, P as heads of PPs, T as head of S (or TP, see forthcoming lecture), etc. Moreover, the syntactic distribution of a word (or its corresponding phrase) is largely determined by its category. Thus, nouns (or rather NPs) are found in subject position, object position, etc., verbs are found in the predicate part of clauses, adjectives modify nouns, adverbs modify verbs, adjectives or other adverbs, etc. The same concerns the syntactic distribution of functional categories, which we shall show later to function as syntactic heads in English. Thus, the distribution of functional category Det is limited to the Noun phrase in English (unlike Arabic where we can find this category in the adjective phrase and the quantifier phrase as well).
Finally, it is worth mentioning that this distinction between words based on their syntactic category may be further supported by empirical evidence provided by language disorders (more specifically the so-called agrammatism or agrammatic aphasia), first language acquisition and second/foreign language learning. In the case of agrammatism, it is well known that agrammatic patients have difficulties only with functional categories, not with lexical categories (though verbs are more problematic for them than nouns). In the case of first language acquisition, it is also well-known that words are acquired following a universally fixed order, depending on their category. Thus, lexical categories, for example, are universally acquired before lexical categories. Finally, the domain of second and foreign language learning can also provide us with solid evidence that humans do distinguish between words based on their category type. For example, the so-called functional categories are recognized to be more problematic for the second/foreign language learner than lexical categories. 2.2 Arguments for phrasal categories In Grammar3 you studied phrasal categories from a descriptive point of view. More specifically, you dealt with five types, namely the Noun Phrase (NP), like the one between brackets in sentence (2a) below, the Verb Phrase (VP), like the one between backets in sentence (2b), the Adjective Phrase (AdjP), like the one between brackets in (2c), the Adverb phrase (AdvP), like the one in (2d), and the prepositional phrase (PP), like the one in sentence (2e):
- a. Please help [this very wise boy] b. Peter will [wash the dishes] c. she is more [intelligent than her brother] d. Peter speaks [very slowly] e. Jane reads books [in the library]
Another argument for the existence of phrasal categories has to do with the well-known syntactic phenomenon or operation of movement, with which we shall deal in the second part of this course. Movement, as we shall show you, is the (universal) syntactic operation whereby specific syntactic constituents (heads, phrases, or clauses) undergo movement or displacement from their original position in the sentence to another position. This is the case for example in wh- questions where a full phrase of a specific type may undergo what is commonly referred to as wh-movement, as in the following sentences:
- a- Which boy will you help? b- How intelligent do you think she is? c- How slowly does Peter speak? d. In which library does Jane read books? In sentence (4a) above, it is a full noun phrase that has undergone movement, in satisfaction of the universal requirement that only full constituents may undergo this operation. In sentence (4b), it is a full Adjective phrase that has moved. The same applies to sentence (4c), where a full adverb phrase has moved, and to sentence ((4d), where a full prepositional phrase has moved to the sentence/clause-initial position. Concerning the verb phrase, it can also undergo movement as a full syntactic constituent (what is generally called VP-proposing), as in the following example:
- Peter will [wash the dishes], and [wash the dishes] he will. Notice that in the second clause of this sentence, the full VP has undergone movement (VP-preposition) in satisfaction of the universal requirement that only full syntactic constituents may undergo movement in human language. Also notice that this argument or constituency test clearly helps us to identify the verb phrase as a constituent consisting of the lexical verb (not the auxiliary verb), complements/objects of the verb if any and the so-called VP adverbials or adjuncts (not disjuncts and conjuncts).
Argument 3: Pro-forms (The Pro-form test) An equally compelling piece of evidence or argument for the existence of phrasal constituents in English (and other human languages) concerns the linguistic phenomenon of proforms. These are single words (such as pronouns) which are used to refer to an already mentioned full constituent. Actually, almost each of the major phrases in English has its corresponding proform, as shown by the following examples:
- a. This very tall girl thinks that she is clever b. Peter will wash the dishes and so will his wife. c. Mary is very helpful and so is her brother d. John will stay in the garden but Peter will not stay there. In sentence (6a), the pro-form 'she' refers to a full NP, namely 'the very tall girl'. In sentence (6b), the proform 'so' refers back to a full VP, namely 'wash the dishes'. In sentence (6c) the proform 'so' refers back to a full AdjP, namely 'very helpful', and finally, in sentence (6d) the proform 'there' refers back to a full PP, namely 'in the garden'. 2.3. Clausal categories. 2.3.1. Definition and classification of clauses Clausal categories are syntactic constituents like the ones put between brackets in the following sentences:
- a. [I think [that John is happy]] b. [The man [whom I saw yesterday] is my friend]. c. [I am doubtful [whether he has understood].
phrase (having phrase internal functions, e.g. direct object, complement of preposition, etc.)) or inside another clause (having a clause internal function, e.g., a subject or disjunct, as in sentence (2a,b) below:
- a. [That the earth is round] is well-known. b. [To be frank with you], you should not leave early. 2.3.2. Arguments for the existence of clausal categories. As already mentioned, clausal categories are well-known constituents of sentence structure. In a forthcoming lecture, we shall deal in detail with their syntactic structure, as viewed by the Principles and Principles framework. In the present lecture, however, next to our definition and review of the traditional classifications of clauses, we shall present few arguments supporting the existence of this (universal) syntactic category. The first argument has to do with coordination. Thus, clauses, like phrases, may be coordinated with one or more similar clauses, as in the following examples (the conjoined clauses are put between brackets):
- a. I think [that you are well-behaved] and [that you deserve my full support] b. John is the person [who works hard] and [who likes great achievements] Since only full constituents may undergo coordination in human language, we conclude that clauses are also full constituents since they can undergo this syntactic operation.
The second argument concerns the constituency test of movement. Again, like phrase, full clauses may undergo movement as well, as in the following example:
- The man has just arrived [whom nobody likes]. In this example, the relative clause (between brackets) has clearly moved from a position inside the complex Subject NP "The man whom nobody likes" to the sentence final position. This movement operation is referred to as clausal extraposition, a common operation in English. For these and other arguments supporting the existence of clausal categories, please read the relevant chapters in your Syntax coursebooks. 3. Further arguments for the hierarchical structure of sentences. The claim we have been defending so far is that the syntactic structure of sentences in English (and other human languages) is hierarchical in nature. The arguments we have presented so far to support this claim have concerned only the so-called constituency tests, that is linguistic tests of various types that confirm the existence of different types of phrasal and clausal constituents that are hierarchically related. Among these tests, we have dealt with coordination, movement, proforms, etc. For more constituency tests (e.g., the deletion test, the clefting test, etc.), please read the relevant chapters in your Syntax course books. Today, we shall present further arguments for the hierarchical nature of syntactic structure. More specifically, we shall deal with four arguments, namely the argument based on the distribution of adverbial expressions in
c. *John can rely on my certainly support. It should be clear therefore that the distribution of adverbials in English clearly show that the syntactic structure of sentences is hierarchical in nature. Otherwise, if the syntactic structure of sentences were flat, that is non-hierarchical (with all constituents being positioned at the same flat level), we would expect all adverbs to have a similar distribution and therefore all sentences with adverbials to be all grammatical. Before moving to the next argument, please note that more details about the argument involving adverbials in Radford's book Transformational Grammar (chapter 2). 3 .2. Subject auxiliary inversion Another important syntactic argument supporting the claim that the syntactic structure of sentences is hierarchical in nature has to do the grammatical operation traditionally referred to by traditional English grammarians as Subject-Auxiliary inversion, an operation which takes place especially in main yes/no questions, as in the following examples (this argument was originally presented in Chomsky (1988), Language and Problems of Knowledge: The Managua Lectures ):
- a. Is John working hard? b. Is that student who is very tall working very hard? What is worth observing here is that when we have more than one auxiliary in the sentence that we want to transform into a yes/no question, like the one in (5b) above, it is not the left- most auxiliary (i.e. the first auxiliary in the sentence) which should be fronted to the sentence-initial position (i.e to undergo subject-aux inversion), but rather it is the highest auxiliary that should undergo this operation, as is the case in (5b) above. If the rule of Subject-aux inversion (or T-to-C movement ) were not formulated in hierarchical terms (i.e. making reference only to linear order, with language having a non-hierarchical structure), we would get ungrammatical yes/no questions like the following:
- *Is that student who very tall is working very hard? 3 .3. Semantic role assignment A further piece of evidence supporting the hierarchical nature of sentence structure concerns the phenomenon of semantic role assignment, i.e., assignment especially to Noun Phrases specific semantic roles (also called thematic roles) such as 'agent', 'patient', 'experiencer'. For example, in the following sentence, the subject NP "John" is assigned the semantic role 'agent', while the object NP is assigned the semantic role 'Patient' (or theme):
- John pushed Mary. As observed by Chomsky (1986), the object NP is assigned its semantic role (patient/theme in sentence (7) above) by the full/lexical verb for which it serves as a complement and which is therefore syntactically represented as a structural sister of this verb, whereas the subject NP is assigned its semantic role (Agent in the same sentence above) not by the verb alone, but by the whole VP, which is the structural sister of the subject. Evidence for this kind of semantic role assignment asymmetry comes from sentences like the following (Chomsky 1986):
- a. John broke the window. b. John broke his leg. In both sentences (8a,b), the object NP is assigned the same semantic role by the verb, namely "patient/theme". By contrast, the subject NP is not assigned the same semantic role in both sentences. In sentence (8a), it is assigned the semantic role "Agent", while in sentence (8b), it is assigned the semantic role "patient/theme", since John is now the victim of the act of breaking his leg. To account for this Chomsky assumes that while the semantic role of the direct object is assigned by the verb alone, the semantic role of the subject is assigned
Thus, our sentence in (10) above has the following two possible syntactic representations:
- a. [S Mary [VP hit [NP the man [PP with red shoes]]]] b. [SMary [VP hit [NPthe man] [PP with red shoes]]]] Notice that in (11a) above the prepositional phrase "with red shoes" is syntactically represented as a constituent of NP, hence it is interpreted as modifying the noun (giving us the first interpretation, i.e., "Mary hit the man who was wearing red shoes"), whereas in (11b) the same prepositional phrase is syntactically represented as a constituent of the VP, hence functioning as a VP adverbial/adjunct (thus giving us the second interpretation, i.e., "Mary hit the man by using red shoes". It should be clear therefore that structural/syntactic ambiguity in human language is direct result of the hierarchical nature of syntactic structure. If sentences had a flat (i.e., non-hierarchical) structure, we would not be able to account for this structural/syntactic ambiguity in a principled manner. Our next lectures will deal with the internal syntactic structure of phrasal categories and clausal categories, as viewed by X-bar Theory, developed within the generative framework (as an alternative to the old (and inadequate) approach of Phrase Structure Grammar (first version of Generative Grammar). Therefore, besides reading the chapters concerning the topics covered so far in this course, you should also read and summarize the relevant chapters of your coursebooks (and other references of your choice)- Best of luck.
