Syntax is the study of the structure of ‘sentences’. Just as with the concept of word, however, – or even more so – there is a problem with the common, intuitive definition of sentences that we may have learnt in school. There, we may well have been told that a sentence is a unit of sense that is constructed of a number of words and bounded by a punctuation mark, such as a full stop, exclamation or question mark. However, a syntactic unit (the term henceforth adopted instead of sentence) needn’t actually contain more than one word in order to fulfil a proper discoursal function. A simple word like yes, no, tomorrow, etc. in answer to a question may constitute a fully sufficient answer. Hence, it depends very much on the function of a syntactic unit whether we could interpret it as the “expression of complete thought” the ancient Greeks postulated as its definition. Before we start talking about functions of syntactic units/categories, though, we’ll first have a look at the potential forms of these units.
Just as with morphemes, with syntactic units we can also distinguish between simpler and more complex forms or types. Following Lyons, we can sub-divide the latter further into compound and complex units. Below, you’ll find lists of all the different types, including examples from the Trainline Corpus.
Look through the lists of unit types above and:
Having done exercise b) above, you will have probably realised that at least part of the function of syntactic units is signalled via their internal structure. In descriptions of this internal stucture, you will often find the expression ‘word order’ used to describe different syntactic configurations. However, this expression is actually quite misleading as words inside a syntactic unit cannot simply be shifted around at random, as the following example program will demonstrate. In order to run this demonstration, simply paste the following units into the text box on the left and click on the scramble sentence button repeatedly to see the result: this is it; the cat ate the rat; the cat sat on the mat.
As you will have noticed, if we use simple (and short) constructions involving pronouns and simple verbs, we may actually get some meaningful, grammatical and well-formed examples, but as soon as we use longer examples that don’t only involve these types of words, we rarely ever achieve such good results. Therefore, it doesn’t quite seem to be the order of the individual words in the unit that actually produces the desired effect, but we may actually first need to combine different words into larger sub-units before we can combine them into something that may make sense from a syntactic point of view. Let’s try this using the same words, but ‘grouped’ together in a specific way, where each | marks the boundary between a unit to be pasted into one of the text boxes below: this | is | it; the cat | ate | the rat; the cat | sat | on the mat.
As you will have noticed, in most cases this works so much better than simply playing around with the order of the individual words. Let us therefore say that it is better to change our original notion of ‘word order’ in a syntactic unit to one of constituent order, reminiscent to that we previously observed for words. Nevertheless, even if we start moving around constituents, we may sometimes get some odd results that may only partly be due to the fact that we are possibly using the wrong structure, i.e. possibly producing something that is not grammatically correct or well-formed.
In order to be able explain this better, we need to develop some further descriptive means of establishing the functionality and meaningfulness of words as part of constituents and within the syntactic order. For this purpose, it is useful to distinguish between the syntagmatic and paradigmatic axes within a syntactic unit. The paradigmatic axis essentially describes the exchangeability of different words in a given (constituent) context, as exemplified in the form below:
The syntagmatic axis describes which words from which word classes or which constituents may be combined with each other in order to form a valid syntactic unit.
Using the examples above + the example i’ll just gonna book from the earlier example list, try to identify what the different terms well-formedness, grammaticality & meaningfulness may refer to.
Just like larger syntactic units, individual syntactic constituents are also characterised by their own internal structure. At the level below the traditional sentence unit, we first find the clause, a syntactic (sub-)unit that contains a subject + a finite verb, i.e. a verb form that is not an infinitive or a participle. Thus, many of the simple units we have discussed above could also be referred to as clauses.
The minimal constituent structure to be identified, and which functions as a separate unit, is the phrase. A phrase may consist of only a single word, but may also be more complex and sometimes even incorporate other phrases or even clauses. Below, you’ll find a (non-exhaustive) list of phrase types:
Analyse the following sample sentences into their individual constituents:
The constituent order of syntactic units is usually represented in one of two ways, either as syntax tree
or as a structure of labelled brackets, where each unit is delimited by a set of matching square brackets [] .
[S[NP my friends][VP[V watched] [NP the film]]] or [S[NP my friendsNP][VP [VwatchedV] [NP the filmNP]VP]S]
In both modes of display, S is usually meant to stand for sentence.
Although conceptionally, drawing trees is not a very problematic thing – provided that one is familiar with the syntactic categories used in a particular theory of syntax and how they may combine –, in practice, it involves a fair bit of planning in order to get the layout right. Although most (generative) theories of syntax that use such phrase-structure trees usually assume rules that start ‘dissecting’ the syntactic unit top-down (e.g. a sentence consists of an NP and a VP, symbolised as S→NP VP, etc.), it generally makes more sense to construct the tree bottom-up by first identifying the PoS for each word, then grouping words into basic phrases, basic phrases into more complex ones, and finally arriving at the top level of the unit.
Draw trees for the sample sentences from the previous exercise. Also try to represent the syntactic structure in labelled brackets.
All constituents fulfil specific roles & functions in a syntactic context. There is no universal agreement about all of these, so the following list needs to be seen as a guideline that is partly open to interpretation, at least with regard to the semantically motivated roles.
Using the sample sentences from the previous exercises again, try to identify the syntactic roles and functions of the individual constituents.
As you will have noticed when you did the exercise regarding well-formedness, grammaticality and meaningfulness, it is rather difficult to distinguish between units that are well-formed because they contain ‘the right constituents in the right places’ and those that are grammatical in all respects. In order to be able to do this properly, we really need to make reference to further concepts that explain the relationship between different constituents some more. We will delay the discussion of some of the more theoretically motivated concepts until the next page, but will discuss the more obvious surface-marked concept of agreement here. Agreement, sometimes also referred to as concord or congruence, refers to the fact that in English, as in many other languages, the relationship between certain constituents is usually marked overtly. The best example for this is subject-verb agreement, where the distinction between singular and plural verb forms is marked in the third person singular by adding an {-s} suffix, as in e.g. he likes music vs. they like music . If we use a (CLAWS 7) tagged version of this example, this will become even clearer he_PPHS1 likes_VVZ music_NN1 vs. they_PPHS2 like_VV0 music_NN1 because the tags make the relationship between singular pronoun (1) and third person singular suffix (Z) and plural pronoun (2) and unmarked verb form (0) more explicit.