Notes (Gibbon)

Sentence syntax is a level of description of language (also known as a level of analysis, a level of representation), characterised by well-defined criteria.

In general terms, the word `syntax' means `structure'; in the study of grammar, however, it is generally restricted to `sentence syntax', sometimes also known as `phrasal syntax'. Some linguists also use terms such as `text syntax' and `word syntax', with the term `syntax' used in its more general meaning. The American poet e.e.cummings writes deploringly of those who worry about `the syntax of spring'.

The study of syntax has a number of goals of domain, method and formulation, three of the basic dimensions of scientific activity; see the Figure (1). These dimensions are described in the following sections.

 
Figure 1:  

 

Domain

• The notion of `grammaticality'.
• The syntactic categories of words (Nouns, Adjectives, Verbs, Adverbs, Articles, Pronouns, Prepositions, Conjunctions, Interjections).
• The syntactic subcategories of words (for example, proper nouns vs. common nouns; transitive verbs vs. intransitive or dual transitive verbs).
• The distribution of words as the smallest units of sentence structure, in respect of:
  • inflection and morphological congruence (agreement),
  • syntactic valency (and its relation to cases, prepositions, word order),
  • semantic selection (one knows that unmarried bachelor is generally a contradiction in terms, and that colourless green ideas sleep furiously is an English sentence, but doesn't make much sense in most contexts),
  • anaphoric reference (pronouns, ellipsis, as in when he arrived, John asked for a coffee, or Ken asked who it was that Mary saw,
  • idiomatic and semantic selection (one holds a speech, cows chew the cud.
• The melodic properties of sentences (sentence prosody: intonation and accentuation).
• The hierarchy of consituents (words and phrases) in sentences.
• Sentences as the largest units structured in terms of words.
• Sentences as the smalles units of text structure.

 

Method

The basic method is concerned with parsing, that is, identifying the parts (constituents) of sentences and their functions. The procedures used in this method differ somewhat, but most procedures include selections of the following:

• Substitution of words in the same context.
• Permutation (changing the order) of words in sentences and examining whether the two sentences are equivalent.
• Deletion of words in sentences to investigate optionality of constituents.
• Insertion of words in sentences to investigate the functional equivalence of words and phrases.
• Interviewing native speakers for their `acceptability judgments'.
• Statistical analysis of the distribution and frequency of words in texts or in speech, and estimation of the probability of occurrence of words.
• Construction of models of the `competence' of native speakers which underlies their `performance'.

 

Formulation

• Informal description by the man in the street about what is right and what is wrong.
• Semi-systematic prescriptive accounts, for instance by school teachers, about `correct' grammar.
• Systematic descriptive accounts, for instance in grammars for foreign language teaching.
• Descriptive theories using symbolic representations.
• Formal theories of language based on logic and algebra, and computer programmes which operationalise these theories, using models based on structure such as:
  • Word sequences (strings, chains).
  • Hierarchical tree graphs.
  • Recursive definitions of complex sentences.
  • Abstract machines.
  • Algebraic grammars for formal languages.
  • Attribute-value logics.
  • Predicate calculus.

Words and rules

Some of the basic ideas underlying the notion of `syntax' as sentence structure are the following:

• Grammar = sentence syntax + inflection + intonation
• ... in grammar books : + word formation + orthography + pronunciation
• Basic category: the WORD (and not just in the beginning).
• Rules: regularities, generalisations, principles of combining parts to wholes, and these wholes to larger wholes.

A particularly interesting type of rule defines the basic property of the linguistic creativity, that is, the ability to produce and understand an unlimited set of sentences.

Recursive rules:

• Jack and Jill are nouns, shouts and screams are verbs; Jack screams is a sentence, Jill shouts is a sentence: a noun concatenated with a verb is thus -- in the simplest case -- a sentence.
• A sentence concatenated with `and' and with another sentence is a sentence: the consequence of this definition is that the resulting sentence concatenated with `and' and yet another sentence is also a sentence, and so on ad infinitum ... Jack screams, Jill screams, Jack shouts, Jill shouts, Jack screams and Jill screams, Jack shouts and Jill screams , ... with this tiny vocabulary, the sentences rapidly become repetitive, but an infinite number of sentences can still be formed.
• Nothing else is a sentence (at least for this somewhat minimalistic grammar).

Generative grammar

The basic idea that language is systematically creative, that is, neither simply a list of boring facts, nor totally chaotic, and that this property is defined in the central component of grammar, was first formulated clearly by Noam Chomsky. The basic general tenets of generative grammar are:

• productivity, creativity,
• innate language ability -- not the details of different languages but the principles of language in general which distinguish human beings from other creatures.
• modularity of human cognition and the autonomy of the language ability, as opposed to derivation of the language ability from very general principles of learning and behaviour.
• an underlying and highly systematic `competence', i.e. language ability, underlies the enormous variations in actual `performance' of utterances in concrete situations.

Definitions of `sentence'

The notion of `sentence' has been given many definitions:

• grammatical string of words,
• string of words which can be described by a grammar,
• an utterance,
• expression of a complete thought,
• expression of a proposition,
• basic unit of text, discourse or argument.

A syntactic poem in German

syntax cycles and boys

 

dafydd gibbon 1992

 

 

der kleine junge sprang auf das rad

auf das rad sprang der kleine junge

sprang der kleine junge auf das rad

sprang auf das rad der kleine junge

auf das rad der kleine junge sprang

der kleine junge auf das rad sprang

der junge kleine auf das rad sprang

kleine der junge auf das rad sprang

kleine junge der auf das rad sprang

kleine junge auf der das rad sprang

kleine junge auf das rad der sprang

kleine junge auf das rad sprang der

der das auf rad junge kleine sprang

auf der das kleine junge rad sprang

The literarily inclined will have noticed that this is a sonnet.

The WORD -- grammar and lexicon

• An older view: A lexicon is just a collection of idiosyncratic and unsystematic properties of words, a kind of linguistic trivial pursuit.
• The contemporary view: The lexicon is at the core of grammar, and sentence structure is determined by systematic properties of words: words have generalisable combinatorial properties which determine the structure of larger units.
• Generalisable properties of words:
  • A `morphological grammar' or `word syntax' determines the construction of new words (word formation).
  • The lexicon contains `attested', `lexicalised' words, i.e. words with attested stems, in addition to the potential words defined by word formation rules.
  • The lexicon indeed also contains idiosyncratic properties of words -- the features which determine how each word is different, to some extent, from other words.
  • However, the similarities between words in the lexicon can be described in terms of hierarchies of interlocking lexical fields, including semantic fields.
• The determination of properties of larger units by properties of their parts is known as projection.
• The features which are projected on to larger units (for instance, if a subject noun is plural, the whole noun phrase, including the article, and the main verb are also plural) are known as head features.

Syntax online links

Several introductions to syntax are available on the web. A selection is provided by Georgetown University. For those interested in computational linguistics, there is a link to a demo of automatic syntactic analysis, with several selections of grammars and lexica, including a facility for entering your own grammar and lexicon.

For those interested in going further in getting additional views on the topic, the AltaVista search engine identifies a wide range of choices for the key phrase Introduction to syntax. Take your pick ...