‘There’s No Such Thing As Syntax –

And It’s a Good Thing, Too...’

Robert de Beaugrande

United Arab Emirates University, Al Ain

 

[COMMENTARY:  this paper was composed for the Festschrift in honour of my dear friend Jan Firbas on his 80th birthday. Firbas sent me an interesting response, which, with his kind permission, I have edited and included here following my own paper.]

 

1. The title of my paper was playfully cobbled from two quotes. Givón (1979: 82) was telling how, for reasons similar to those I shall air, he arrived at the view that ‘syntax per se does not exist at all’, but he later ‘retreated’ from it. Fish (1994) was demystifying the exploitation of ‘free speech’ to protect the purveyors of sexism, racism, and ethnic prejudice. He pointed out how ‘speech’ is never ‘free’, and certainly not when it is driven by an ideology as ham-fisted and yet ham-strung as white male supremacy. My own argument will be that ‘syntax’ as the term is generally purveyed in ‘theoretical linguistics’ does not meaningfully apply to real language. Of course language does have a syntactic dimension in some broad functional sense, but not a any separate ‘level’ or ‘component’ of ‘syntax’ in a narrow formalist sense.

 

A. Seven senses of ‘syntax’ in the drift of ‘formalism’

 

2. I shall start by enumerating some seven senses of the term ‘syntax’ extracted from my extended inquiry into the discourses of influential linguists. To conserve space, I would refer to my previous volume (Beaugrande 1991) for further details and more specific references.

 

2.1 ‘Syntax’ is the order of words in a temporal sequence of speech. A fundamental step in modern linguistics was to declare the primacy of speech over writing (e.g. Saussure 1916; Bloomfield 1933). Yet foundational linguistic works on ‘syntax’ (e.g. Harris 1951; Chomsky 1957) were expounded using samples of written language, even while consistently invoking the ‘speaker’ and the ‘hearer’ rather than the ‘writer’ and the ‘reader’. The chief motive might be that ‘syntax’ is harder to identify within a temporal continuum of spoken sounds than within a spatial sequence of discrete written words. Besides, written language had been the focus in centuries of disquisitions commonly subsumed under ‘traditional grammar’.

2.2 ‘Syntax’ is the order of words in a spatial sequence of writing. For the motives just cited, ‘syntax’ has been tacitly interpreted as a factor of written language. Writing conveniently suspends those features of spoken language that syntacticians find inconvenient to represent, let alone explain, such as articulation, audition, or intonation. Also, a written sentence bears a visual and spatial resemblance to its ‘formalisation’, such as a ‘tree diagram’ with ‘left branches’ and ‘right branches’.

2.3 ‘Syntax’ is the formal arrangement of the sentence. As shown elsewhere, the ‘sentence’ has enjoyed a remarkable career as the central unit of language among linguists of quite different approaches or theories (Beaugrande 1999). The leading motive there might be that the term ‘sentence’ has over the years accumulated a wide range of meanings, including: the ‘expression of a complete thought’; a unit having a ‘subject and predicate’; an orthographic sequence ending with a ‘period’; and a theoretical unit corresponding to the ‘utterance’ as a practical unit. These disparate meanings have allowed linguists to describe all sorts of things in terms of the ‘sentence’, and to declare the sentence-boundary to be the outer limit of their investigations (cf. § 13, 17). Also, the sentence nicely abetted the focus upon written language over spoken.

2.4 ‘Syntax’ defines a ‘language’ to be a ‘set of sentences’. This sense is a radical reformulation of sense 2.3, and limits the whole of ‘language’, and thereby the scope of linguistic investigation, to the properties of sentences. Moreover, ‘syntax’ has been narrowly defined as an account of how the ‘sentences’ in that ‘set’ are formed from a shared base of ‘rules’. A ‘transformational theory’ simply devises the ‘rules’ for ‘rewriting’ one type of sentence as another, and claims that doing so ‘explains language’. In such a theory, ‘syntax’ could be declared ‘autonomous’ and ‘independent of meaning’ (Chomsky 1957) of a theory of meaning, on the grounds that the ‘meaning’ of a ‘sentence’ is ‘preserved’ during ‘transformation’.

2.5 ‘Syntax’ is the equivalent of ‘grammar’. This sense has usually been implied in formalist work simply by using the terms ‘syntax’ and ‘grammar’ interchangeably. Made explicit, the equation raises problems if ‘grammar’ is held to be separated from the ‘lexicon’ (but cf. § 18). Through that separation, syntax could be studied as the connection of linguistic units without studying the units themselves, the so-called ‘lexemes’. In ‘transformational theory’, the ‘transformations’ could ostensibly preserve their lexical identity as well as their meaning.

2.6 ‘Syntax’ is the ‘generative component’ of a ‘language’. This sense became current to broaden the narrow ‘transformational’ approach by making some provision for meaning in the word-order of the ‘sentence’. So ‘semantics’ was added as an ‘interpretive component’ applied to sentences already put in order by the ‘syntax’ as the ‘generative component’. Significantly, the term ‘generate’ itself did not mean ‘put the words of the sentence in order’ but rather ‘assign a structural description to the sentence’ after it had been put in order — by whom, how, or why were no concerns of the ‘syntax’ (cf. § 19, 37).

2.7 ‘Syntax’ co-ordinates ‘surface structure’ with ‘deep structure’. This sense too attempted to broaden the narrower approaches. The pattern of the written sentence was now renamed ‘surface structure’, whilst the more important ‘underlying’ pattern was the ‘deep structure’. The ‘transformational rules’ no longer just rewrote one type of sentence into another but ‘derived’ this ‘surface structure’ by ‘mapping’ it from the ‘deep structure’, and in the process ‘determined its semantic interpretation’ (Chomsky 1965). This too was heralded as an ‘explanation’ of language, indeed one whose ‘adequacy’ far exceeded any other ‘linguistic theory’. But in effect, the new design vastly diminished the interest of ‘syntax’ in analysing the order of words in sentences. In return, the categories of ‘deep structure’ were claimed to be ‘universal’ and thus to ‘explain language acquisition’ as the categories applied by the child’s ‘innate language acquisition device’ (see Beaugrande 1997b, 1998a for discussion).

 

3. This, then, is a list of seven major senses of ‘syntax’ entailed in linguistic discourse. They demarcate a steady increase in ‘theoreticalness’ to render the scope of research steadily more abstract and ‘formalist’. Spoken language was displaced by written; a ‘language’ was reduced to a ‘set of sentences’; ‘surface structure’ was marginalised by ‘deep structure’. Eventually, even ‘sentences’ faded away, and ‘syntax’ was relocated in the former terrain of ‘semantics’ as the ‘study of the structure of mental representations’ and of ‘the relation between words and concepts’ (Chomsky 1991: 93f).

4. To say ‘there’s no such thing as syntax’ would mean that the term in these seven senses does not cover any aspect of real language. Real speakers put words in order for many reasons which this ‘formal syntax’ does not address and which cannot be stated in ‘formal rules’’ transformational or otherwise. Moreover, this ‘syntax’ has progressively emptied out and idealised ‘language’ itself until that term also no longer refers to human language in the real world. Unique among modern sciences, such a linguistics replaces its natural object of investigation with one of its own invention.

5. Why might we add ‘it’s a good thing too’? If one set of formal rules did determine the word-order of ‘all grammatical sentences’ and of no ‘ungrammatical’ ones independently of meaning and lexicon, the learning and use of language would be unmanageably complex. This implied complexity was masked by postulating a ‘language acquisition device’ to manage it all and by marginalising language use as ‘surface structure’, ‘performance’, ‘e-language’ and so on. Formalist linguistics then introduced its own arbitrary and gratuitous complexity quite unrelated to the strategic and operational complexity of any real language (Beaugrande 1998a).

6. We might decide to discard the term ‘syntax’, but I doubt whether this could be done. So our recourse would rather be to redefine the term and re-integrate it with the multiple modes of word-order in real language.

 

B. ‘Multifunctional syntax’

 

7. This redefinition has in fact long been on the ‘functionalist’ agenda of the ‘Prague School’. I once found that name a bit odd, since some key figures, including Jan Firbas, were not in Prague or only temporarily, and since the school was long disapproved by the occupational regimes ruling from that very city. But perhaps the city-name stuck just because the school was hard to describe in the terms of conventional linguistics. The Prague School work was rich and diverse far beyond most of ‘Western’ linguistics, but, aside from phonology, was long resisted by political repression and academic compartmentalisation (Beaugrande 1997c). And that resistance was partly due to its broad conceptions of ‘syntax’.

8. In Prague School work, the term ‘functional’ has been thematic. It figured in the English title of Mathesius’s Functional Analysis of Present Day English on a General Linguistic Basis, a volume not published in Czech until 1961 and in English until 1975, but essentially completed in manuscript around 1935 and circulated thereafter among his students and colleagues. As I read that book, his conception of ‘functionalism’ — witness his own programmatic term ‘linguistic characterology’ (Mathesius 1928) — sought to integrate the comparative and the historical dimensions of languages, in line with established ‘philology’. So his own ‘functional analysis’ (obsahovv rozbor) explored how the organisation of ‘present-day English’ compares to that of languages like Czech, German, and French, and how it historically evolved through Old and Middle English. This ‘general linguistic basis’ was the total opposite of Saussurian ‘general linguistics’, which disregarded both the comparative and the historical dimensions of language.

9. The ‘functional analysis’ expounded by Mathesius quite naturally led to the a ‘functional sentence perspective’ (FSP)1 defined for English by native speakers of Czech. On the comparative dimension, these linguists applied the perspective of their native language, with its rich ‘functional’ constraints upon word order, over onto English and discovered similar constraints which native speakers of English have rarely recognised (see now Daneš 1995). On the historical dimension, they found the ‘functional’ constraints applying more firmly to the word order of Old English and Middle English closer than Modern English (Mathesius 1907; Šimko 1957; see now Firbas 1992). Such insights were hardly welcomed to the non-comparative and non-historical linguistics in the ‘West’.

10. Still, their view of syntax was not just functional but multifunctional. In his far-sighted paper on ‘comparative word order studies’, Mathesius (1942) proposed the ‘grammatical principle’, the ‘FSP principle’, the ‘emphasis principle’, and the ‘rhythmical principle’. The concept of a ‘principle’ implies the synthesis of parallel and multiple applications to the same sentence or utterance, as opposed to parcelling out an analysis into distinct ‘levels’ or ‘components’ in the manner of formalism. These ‘principles’ are accordingly not separated: all four can help determine the actual order words. In a language like Czech, the ‘grammatical principle’ predictably follows the ‘FSP principle’ more closely than in a language like English. Czech compensates for variations in word-order with formal inflections indicating grammatical functions. The grammar of English, in contrast, lacks inflection and relies in ‘frozen islands’ of apparently fixed patterns. This factor explains why English has been a favourite language of the formalists who saw ‘syntax’ as ‘autonomous’ and ‘independent’ (§ 2.4).

11. Later, Skalička (1960a, 1960b) proposed a signally multifunctional ‘syntax’ with two slightly different schemes, including: the syntax of FSP; the ‘implicit’ syntax of the lexicon; formal syntax (rhyme, alliteration etc.); stylistic syntax; and the ‘syntax of sentence connection’. Here, the term ‘syntax’ was significantly expanded to provide for FSP and also to incorporate the contributions of the lexicon to word-order, and to look beyond the boundary of the individual sentence (cf. § 15).

12. Shortly after Skalička, Daneš (1964) proposed a ‘three-level approach to syntax’ which ‘distinguished’ ‘the sentence’ (1) as ‘a singular and individual’ ‘utterance-event’; (2) as a ‘minimal communicative unit (utterance) of a given language’; and (3) ‘as an abstract structure or configuration’ within the ‘grammatical system of a given language’ (1964: 229). These three ‘levels’ could correspond to three steps in the linguist’s analysis. The first step would address the ‘utterance-event’ of ‘speech (la parole)’ as it is ‘immediately accessible to our observation’. The second step would set about ‘depriving such an event, by way of abstraction, of all accidental, singular, and individual elements’ so as to attain ‘an utterance which no longer belongs to speech’ and has ‘non-grammatical but systemic means of organisation such as word order’ and ‘intonation’, including ‘many more features than only those belonging to the most abstract and general syntactic pattern of the grammatical system’. Here, the ‘sentence’ figures as an ‘utterance’ that ‘remains a part of context and situation’, and ‘contains concrete lexical items and elements of modality’ and ‘emphasis’ (§ 16, 21, 31). The third and ‘highest step of generalisation’ would yield the ‘specific grammatical’ ‘sentence pattern’, which ‘represents an abstract and static invariant structure (scheme), not a sequence of particular words’ within an ‘utterance based on this underlying pattern’ (1964: 230f).

13. The contrast here with the formalist ‘syntax’ described in section A is striking. The typical formalist skips the first step altogether and performs the second step quickly and quietly, ‘depriving the event’ also of all ‘non-grammatical organisation’ and ‘intonation’, and leaving only ‘the most abstract and general syntactic pattern’. Nearly all of the analysis treats what would be the third step in Daneš’ scheme as if it were the first and only step, thereby casually obscuring a deal of significant data about real language. In particular, the individual predispositions of the linguist are tacitly introduced in the guise of the ‘ideal speaker-hearer’ and a purely theoretical informant.

14. We should also note that the ‘model’ proposed by Daneš invokes and yet reconciles the staid and unproductive dichotomy between ‘langue and parole’ or ‘competence and performance’ which has justified the formalists in setting aside any aspects of real language they found hard to formalise. Daneš proposed instead a linguistics that expressly starts with the ‘individual utterance-event’ and subjects it to a carefully controlled process of ‘abstraction’ and ‘generalisation’. In a kindred vein, Firbas (1962) proposed to take a fresh look at the ‘function of the sentence in the act of communication’.

 

C. ‘Multifunctionality’ in large-corpus data

 

15. To build up these multifunctional approaches from the ‘Prague School’, we can devote more focussed research to Daneš’ first and second steps. An alternative method to transcend what the ‘singular and individual’ nature of the ‘utterance-event’ would be to compare a large number of similar events, using the authentic data in a very large corpus. We can also transcend the limitations to the single ‘sentence’, which were all along less vital for functionalism than for formalism. We can instead look into the functional motivations of speakers or writers for formatting any ‘utterance-event’ as a sentence (cf. Beaugrande 1999).

16. Large-corpus data may implicate some ‘missing links’ between language and text (or ‘langue and parole’, or ‘competence and performance’) (Beaugrande 2000). The data display the ‘competence’ of the language community bent far toward ‘performance’, and the community’s ‘performance’ bent far toward its ‘competence’. The larger the corpus, the clearer our vision.. In the words of Daneš quoted in § 12, we can encounter ‘systemic means of organisation such as word order’, and can examine a ‘sentence’ or ‘utterance’ that ‘remains a part of context and situation’ and ‘contains concrete lexical items and elements of modality’ (cf. § 21).

17. This work can highlight the interactions among sets of multifunctional choices. In the ‘lexicon’, these interactions constitute the collocability in the language, and the actualisations in a texts are the lexical collocations. In the ‘grammar’, these interactions constitute the colligability in the language, and the actualisations in texts are the grammatical colligations. Obviously, no one set of collocations and colligations, however large, could display the total collocability and colligability for any real language. But enlarging the set does enable steadily more delicate descriptions of the criteria of selection and combination (Sinclair 1998). These criteria are not readily open to unaided intuition, and so cannot be extracted from the small sets of invented ‘sentences’ routinely described in formalist linguistics (cf. Francis and Sinclair 1994). Also, the criteria help us reconstruct how a sentence or utterance might have been put in order, and by whom, and why, instead of just ‘assigning it a structural description’ after the fact (cf. § 2.6).

18. Interestingly, collocations and colligations often cluster together, reflecting the integration of ‘lexicon’ and ‘grammar’ within the ‘lexicogrammar’ (cf. § 2.5). Here, vast work remains to be done, and our conclusions can only be speculative for the present.

19. My data sample was taken in July 1994 from the Collins Birmingham University International Language Database (COBUILD), also affectionately called the ‘Bank of English’. It was (and still is) the world’s largest computerised data corpus, then 200 million words of running text, of which 20 million were in spoken language. The proportions of spoken data to might be much larger to match those of ordinary language experience, but a data sample of 20 million words is incontestably impressive and frees us from the formalist move of continually invoking the ‘speaker’ and the ‘hearer’ whilst relying solely upon written language (§ 2.1).

20. My data sample centres on the English target expression ‘saw to it that’. As a collocation, it prefers specific kinds of Agents both as the Subject of ‘see to’ and the Subject of the Clause telling what got ‘seen to’. As a colligation, it allows for grammatical variations only in Person and Tense; we would not find Passive Transitivity (‘*they/the event was seen to that’), nor other Pronoun Objects (e.g. ‘*saw to you/her/him that’), nor again Adverbs inserted after the Verb (e.g. ‘?saw resolutely to it that’).2 A more comprehensive study would require queries for all Tenses; but the present paper will summarise only in the data lines returned for the Simple Past.

21. As often with corpus data, the total meaning of the expression is not the sum of word-meanings. I found virtually no data where the Verb ‘see’ had the meaning listed in first place in a typical dictionary, e.g. ‘to perceive with the eye; to view’ (Random House Webster’s, p. 1213). Instead, the meaning of ‘see to it that’ is roughly ‘create the conditions needed so that’ or ‘make it necessary that’, where the ‘elements of modality’ in the sense of Daneš (§ 12) are fairly subtle as compared, say, to ‘commanded that’.

22. Nor are we likely to find the corresponding Verbs for ‘see’ in other languages being used this way. A query of the German data-base of the Institut für deutsche Sprache in Mannheim3 returned no such uses of the German Verb ‘sehen’. One would need to use expressions like ‘veranlassen’, in more formal usage, and ‘dafür sorgen’ or ‘es dazu bringen’ in more informal usage. So learners of English as a non-native language would not readily recognise or predict its range of collocations and colligations (cf. § 34).

23. I shall first summarise the preferences for putting the words in order before the target expression. The Subjects were typically Agents of power, as in (1-7).

(1) The state saw to it that they were issued a work permit, fingerprinted, photographed

(2) The personal intervention of the Führer saw to it that the Gestapo prevailed

(3) British Prime Minister Winston Churchill saw to it that Murrow had access to all that he needed

(4) the World Bank […] saw to it that the plant got appropriate supplies of power

(5) Zeus thought better of the relationship and withdrew and saw to it that she should marry a human husband

(6) Divine Providence, always fair when interfering in human fate, saw to it that all ended well

(7) God saw to it that someone survived to tell this story

By implication, such Agents or their subordinates used any necessary means to obtain the result. The greater the power, the less the need to explain the means, e.g., how Hitler and Churchill proceeded, not to mention Zeus, ‘Divine Providence’, and God.

24. As a more ordinary variation, the Agent of power was a parent or family member deciding things for children, as in (8-11).

(8) Sonia saw to it that her family always looked smart for church

(9) in the morning, standing erect and still, she saw to it that they washed behind their ears, inside their nostrils

(10) His mother hoped for a time that he would become a priest and saw to it that he became an altar boy at seven

(11) His father now tried to interest him in the opposite sex, and saw to it that the palace had frequent visits from eligible princesses

Here too, the means can remain unexplained, e.g., what the parents did to make the children ‘look smart for church’ and to discourage them from walking abroad with ‘unwashed nostrils’ (who’d ever notice?).

25. Next, I shall summarise the preferences for putting words in order after the target expression. The Subjects of the next Verb were persons in 14 data samples, e.g. (3) and (8-10), and (12); institutions in 10 more, as in (2), (4), and (13-14); and things associated with institutions in another 13, as in (11) and (15-16).

(12) The Depression and the Wall Street crash saw to it that Stella never got anything like what was intended

(13) local political interests saw to it that Congress heard and responded to their complaints

(14) The agreement saw to it that local telephone exchanges served a rather limited area

(15) He saw to it that policy decisions, once taken, were given the means

(16) Lane finally saw to it that the appropriation for the hospital was raised

Such data again conveyed implications of power exerted by authoritative Agents upon Events or other Agents.

26. A similar factor might account for the predominance of colligations with Passive Transitivity in 24 data samples, as in (1) and (15-16), over 15 with the Active, as in (9) and (14), and over 12 with the Medial, as in (6) and (10). Also, a number of ‘quasi-Passive’ collocations, though were technically not Passives still indicated something getting done to the Subject, such as ‘got supplies’ (4) and ‘had and ‘had visits’ (11). Here too, the implication would be the imposition of one Agency upon another. Also, Passives and ‘quasi-Passives’ are well-suited when the direct Agent or means is not specified, as in (4), (8), and (15-16).

27. Another significant colligation was the frequent Thematic Subjects in the sense of FSP. The immediately following words included the Definite Article ‘the’ in 16 samples; and the Pronouns ‘they’ in 3, ‘he’ in 2 plus ‘his’ in 4, and ‘she’ in 2 plus ‘her in 1. Perhaps the target expression tends to be used when the context is already prepared, e.g., when we know which ‘plant’ (4) or ‘palace (11) is meant, and who co-refers with the Pronouns. The ‘it’ in ‘saw to it’ has no lexical meaning and occupies a position which is vaguely cataphoric as contrasted with Noun Objects, as in ‘see to her comfort’ or see to our needs’.

28. Even the Rhematic Subjects with Indefinite Article in my data seemed predictable in hindsight:

(17) we had the letter typed at the dacha; some helpful assistant saw to it that a copy was on Kosygin’s desk the very next day

(18) knowing of Mountbatten’s wishes that what he had said should be known, I saw to it that a transcript of the Suez programme reached my friend Bernard

A ‘typed letter’ would lead to ‘a copy’ (17); and what ‘Mountbatten had said’ could be effectively made ‘known’ by ‘a transcript’ (18).

29. In a similar manner, I found many instances of ‘lexical cohesion’ in a somewhat broader sense than Halliday and Hasan (1976), namely when the Subject after ‘saw to it that’ was associated with some previous lexical expressions. Samples (19-21) logically had Thematic Subjects with Definite Article; one sample showed a recurrence of the same lexical Word-Stem (21). Also, one Rhematic Subject with Indefinite Article was also anticipated this way (22).

(19) At sea, the possibilities were more limited, but Fruhling saw to it that even the smallest vessels of the fleet

(20) in a jewellery store and saw to it that the ring was purchased

(21) the better-off, who themselves were inadequately housed, rapidly saw to it that the housing was reallocated

(22) She taught her music and painting and saw to it that every week a long composition was written

These data might illustrate what Skalička called the ‘implicit syntax of the lexicon’ (§ 11): words associated in meaning or topic appearing fairly close together in syntactic sequences.

30. The corpus data I have briefly surveyed manifest a typical paradox. Every sample is unique at some degrees of delicacy our intuition could not predict or invent, yet most seem intuitively quite acceptable when placed before us (§ 17). The regularities of collocation and colligation are apparently not so much on the ‘surface’ of actual words as somewhat ‘below the surface’ in the dynamically evolving criteria of selection and combination. Thus, we find a strong preference for ‘saw to it that’ to have an Agency of higher power as its own Subject and to have an Agency of lower power as the Subject of the following clause. Since ascribing power to oneself might seem pompous, I found only two samples with a Subject in the First Person, as in (18); the other one reversed the tone by deprecating the writer:

(23) to be among poets, novelists, playwrights, and all manner of belletrists, I therefore saw to it that my corduroys were suitably wrinkled, my suede shoes scuffed

Avoiding pomposity might also be a reason why the Second Person was not found at all.

31. We are accordingly able to confirm the ‘systemic means of organisation’ Daneš envisioned for word order’, provided our language data ‘remain a part of context and situation’ (cf. § 12). If ‘syntax’ gets disconnected from context and described strictly as a formal sequence of abstract categories, that systemic order disappears, and ‘parole’ or ‘performance’ is viewed as a mass of disorder (Beaugrande 1997a, 1998a). The formalists preferred to invent their own modes of ideal order such as ‘deep structure’, which replace the real complexity of language with an artificial complexity (§ 5).

32. The ‘functionalists’ (or ‘multifunctionalists’) of the Prague School, in contrast, never performed this portentous disconnection. They based their own enterprise firmly upon the comparative and the historical dimensions of philology (§ 8f). They did not strike some programmatically ‘scientific’ posture, e.g., by abjuring the study of meaning as Bloomfield (1933), Harris (1951), and Chomsky (1957) did. In return, their own conception of language was too rich and deep to be systematically elaborated until the advent of very large samplings of actual data made accessible by corpora.

33. As the new millennium commences, work with large-corpus data portend a profound reassessment of ‘linguistic science’. We have long dwelt upon the ‘abstract and static invariant structures’ reached by the ‘highest step of generalisation’ envisioned by Daneš (§ 12). We can now explore a new order of concrete dynamic regularities emerging in the collocations and colligations of authentic data. Contrary from what a ‘universalist’ linguistics would claim (2.7), such regularities are far from trivial for understanding the nature of language. They show how that the immense complexity of language is continually integrated during discourse, such that language must have a strategic design for precisely this mode of operation.

34. Moreover, such regularities are highly informative for learners of English as a non-native language (§ 22). Issues of power and authority are largely regulated by language and discourse, and learners should be made aware lest they say something untoward like ‘I saw to it that you would be here’. A key step might be empowering learners to browse authentic data during their own processes of language learning and discovering the regularities for themselves (Beaugrande 1998b). They would then be strategically trained to detect regularities in other samples of the English, as opposed to ‘drilling’ the ‘rules’ in ‘grammar lessons’ and memorising lists of words in ‘vocabulary lessons’.

35. Back at the start, I cited Fish’s point that ‘speech’ is never ‘free’ (§ 1). What people say is richly guided by what is typical and expected. Yet corpus data display a inspiring margin of freedom in the collocations and colligations — last-minute on-line decisions about what to select and combine. We may finally establish the merited respect for the multifunctional creativity of ordinary discourse in human affairs.

 

Notes

1 The English term ‘functional’ was a free rendering of Czech term ‘aktual’, i.e., ‘ongoing’ or ‘dynamic’ (cf. Beaugrande 1992).

2 The grammatical terms, written here in Upper Case, follow the ‘functional lexicogrammar’ expounded in full in Beaugrande (1997a: Ch IV).

3 I am indebted to Heiko Schoberwalter of IDS for this query.

 

References

 

Beaugrande, R. de. 1991. Linguistic Theory: The Discourse of Fundamental Works. London: Longman.

Beaugrande R. de. 1992. The heritage of functional sentence perspective from the standpoint of text linguistics. Linguistica Pragiensa 34(1-2): 2-26 and 55-86.

Beaugrande R. de. 1997a. New Foundations for a Science of Text and Discourse. Greenwich, CT: Ablex.

Beaugrande R. de. 1997b. Theory and practice in applied linguistics: Disconnection, conflict, or dialectic? Applied Linguistics 18: 279-313.

Beaugrande R. de. 1997c. On history and historicity in modern linguistics: Formalism versus functionalism revisited. Functions of Language 4: 169-213

Beaugrande R. de. 1998a. Performative speech acts in linguistic theory: The rationality of Noam Chomsky. Journal of Pragmatics 29: 1-39.

Beaugrande R. de. 1998b. Society, education, linguistics, and language: Inclusion and exclusion in theory and practice. Linguistics and Education 9(2): 99-158.

Beaugrande R. de. 1999. Sentence first, verdict afterwards: On the long career of the sentence. Word 50: 1-31.

Beaugrande R. de. 2000. Text linguistics at the millennium: Corpus data and missing links. Text 20.

Bloomfield, L. 1933. Language. New York: Holt.

Chomsky, N. 1957. Syntactic Structures. The Hague: Mouton.

Chomsky, N. 1965. Aspects of the Theory of Syntax. Cambridge: MIT Press.

Chomsky, N. 1991. Language, politics, and composition. In Interviews: Cross-Disciplinary Perspectives on Rhetoric and Literacy, G. Olsen and I. Gales (eds.), 61-95. Carbondale: Southern Illinois UP.

Daneš, F. 1964. A three-level approach to syntax. Travaux Linguistiques de Prague 1: 225-240.

Daneš F. 1995. Languages and the science of language in the flux of our epoch. Folia Linguistica 29: 187-193.

Firbas, J. 1962. Notes on the function of the sentence in the act of communication. Sborník Prací Filosofické Fakulty BrnŤnské Univerzity A10: 134-148.

Firbas, J. 1992. Functional Sentence Perspective in Written and Spoken Communication. Cambridge: Cambridge UP.

Fish, S.E. 1994. There is No Such Thing as Free Speech — And It’s a Good Thing Too. Oxford: Oxford University Press.

Francis, G. and Sinclair, J.McH. 1994. ‘I bet he drinks Carling Black Label’: A riposte to Owen on corpus grammar. Applied Linguistics 15: 190-200.

Givón, T. (1979: 82) From discourse to syntax: Grammar as a processing strategy. In Syntax and Semantics 12: Discourse and Syntax, T. Givón (ed.), 81-112. New York: Academic.

Halliday, M. and Hasan, R. 1976. Cohesion in English. London: Longman.

Harris, Z.S. 1951. Methods in Structural Linguistics. Chicago: U of Chicago P.

Mathesius, V. 1907. Studie k dějinám anglického slovosledu. Vzĕstník České Akademie 16: 261-75.

Mathesius, V. 1928. On linguistic characterology. Actes du Ier Congrčs International des Linguistes, 56-63.

Mathesius, V. 1942. Ze srovnávacích studií slovosledných. Časopis pro moderní filologii 28: 180-191.

Mathesius, V. 1961 [1935]. Obsahový rozbor současné angličtiny na základé obecné lingvistikém. Prague: Academia Věd.

Mathesius, V. 1975 [1961]. A Functional Analysis of Present Day English on a General Linguistic Basis. Prague: Academia Věd.

Šimko, J. 1957. Word Order in the Winchester Manuscript and in William Caxton’s Edition of Thomas Mallory’s Morte d’Arthur (1485). Halle/Saale: Martin-Luther-Universität thesis.

Sinclair, J.McH. 1998. Large corpus research and foreign language teaching. In Language Policy and Language Education in Emerging Nations: Focus on Slovenia and Croatia, R. de Beaugrande, M, Grosman, and B. Seidlhofer (eds.), 79-86. Stamford, CT: Ablex.

Skalička, V. 1960a. Syntax promluvy (enunciace). Slovo a slovesnost 21: 241-249.

Skalička, V. 1960b. Über die besonderen Formen der Syntax. Rusko-české studie: Sborník Vysoké Školy Pedagogické v Praze, 37-42.

 

Notes on some basic concepts of the theory of FSP:
With a response to Robert de Beaugrande

 

Jan Firbas

 

1. Functional sentence perspective as the interplay of multiple factors

 

I cannot but agree with Robert de Beaugrande that real language does not display a level of syntax that can be conceived of in a narrow formalist sense. Syntax is an organizing force which cannot be severed from meaning. As for the relationship between meaning and syntax, I should like to point out the following. With an amplification to be stated presently, I accept Mathesius’ view that a sentence comes into existence through the process of naming (the onomatological process) and the process of syntactic structuration. In this connection I find the formulations of Reichling and Daneš (cf. Firbas 1992, FSP in Written and Spoken Communication, p. 14) truly felicitous.

In answering the question whether the functional perspective of the sentence is imposed upon the semantic and syntactic (grammatical) sentence structure after this structure has been created, I find that the language users produce the perspective during the sentence creation process. The communicative purpose to be fulfilled and to be communicated by the perspective has been a concern from the very start.

My enquiries have shown that functional sentence perspective (FSP) is determined by the interplay among multiple factors, each yielding its own signals by its own criteria. In written language, these include the contextual factor, the semantic factor, the factor of linear modification.. In spoken language these three factors are joined with the factor of intonation.

The enquiries have further shown that this interplay constitutes a system, not a rigidly closed one, because it has its periphery; it is, however, a system that is strongly operative. The contextual factor operates by assessing the retrievability of a piece of information from the immediately relevant context, verbal and/or situational. The criteria for this factor are determined by the presence or absence of a piece of information in the immediately relevant context and its re-expression in the act of communication. The criteria yielded by the semantic factor are  determined by the semantic content of a linguistic element and the character of semantic patterns into which this content enters. The criteria yielded by the factor of linear modification are  determined by the positions in the linear arrangement The criteria yielded by the factor of intonation  are determined by the prosodic features at varying degrees of prominence. The interplay of all these factors determines the distribution of degrees of communicative dynamism (CD), which in turn determines the functional perspective of the sentence.

These factors operate in a hierarchical relationship. The contextual factor is hierarchically superior both to the semantic factor and the factor of linear modification. The latter two can fully assert themselves only through the contextual factor. The semantic factor is hierarchically superior to the factor of linear modification. The latter can fully assert itself only through the contextual factor and the semantic factor. As for the operation of the factor of intonation, the distribution of the features of varying prominence, i.e. different degrees of prosodic prominence (PP) either perfectly corresponds to the distribution of degrees of communicative dynamism as determined by the interplay of the non-prosodic factors or deviates from a perfect correspondence between the two distributions. These deviations are functional. They can produce prosodic intensification and raise the degrees of CD.

Enquiries conducted on an empirical basis have gradually established the laws of the interplay among FSP factors and presented FSP as a system. Methodologically speaking, the enquiries have ultimately led to an account of FSP phenomena in the same manner as Mathesius had accounted for the facts of word order. Mathesius actually viewed word order as a system determined by an interplay of word order principles (or factors). Like the FSP factors, the word order principles are hierarchically organized; they are dominated by a leading principle (or factor).

 Mathesius’ account of word order has not yet been fully appreciated. We should recognise that it is applicable to different languages and provides a basis for comparisons resulting in what he called ‘characterologies’: the assessments of typical word order properties characteristic of different languages.

Mathesius regarded FSP as a formative force asserting itself through word order and intonation. He dis not take other factors into account. His article presenting his systemic conception of word order was one of his last papers published before his death (“Ze srovnávacích studií slovosledných“ [From comparative word-order studies], Časopis pro moderní filologii, 28/1942, 180-191]. It is difficult to say if, in his later research, he would have discovered that his insight into word order can, with due alterations, be also applied to FSP. My own enquiries on an empirical basis have revealed that the systemic character of FSP is accountable in terms parallel to those applied to Mathesius’ system of word order.

The degree of CD is the relative extent to which a linguistic element of any rank contributes towards the further development of the communication. The distributional field of degrees of CD determines not only the functional perspective a sentence, but also the functional perspectives of subordinate clauses, semi-clauses (non-finite clauses), and noun phrases, which all serve as distributional subfields. For a detailed discussion of the interplay of factors determining the distribution of degrees of CD, and the hierarchy of carriers of CD and the hierarchy of distributional fields, see Firbas 1992, FSP in Written and Spoken Communication.

In my papers and my book on FSP I have discussed the relations between the factors/principles within their systems and also dealt with the question of the relationship between the system of FSP and that of word order. FSP operates in and affects word order through linear modification. Linear modification asserts itself in word order as the FSP linearity principle.

But due to the interplay of multiple factors, one and the same sentence structure can function in different perspectives in accordance with changing contextual conditions. I would speak of different possible contextual applications of the structure, all its possible contextual applications constituting its contextual applicability. 

If linguists decontextualize a semantic and syntactic (grammatical) sentence structure, they do so, they deprive the sentence of its function in a particular contextual situation. It does not serve any communicative purpose; it does not operate in the act of communication and does not display a functional perspective; the interplay among the factors determining the perspective has been deactivated. A decontextualized sentence structure is a static phenomenon. My own propsal would rather be to ‘deintuicize’ (an expression coined for me by Frank Plewes, an American colleague) our interpretation of a sentence.

If employed in context, the sentence structure takes part in the development of the communication and constitutes a dynamic fact. In the act of communication the sentence structure serves as a communicative field, its syntactic constituents serving as communicative units within it. (I owe the concept of communicative unit to Aleš Svoboda.) The syntactic constituents continue to perform their syntactic functions, but in regard to the development of the communication they perform specific dynamic semantic functions.

How have I arrived at the dynamic semantic functions? In determining the distribution of degrees of CD over the syntactic constituents (communicative units) of the sentence structure, I found that the sentence is ‘perspectivised’ either towards the subject or away from it. In the former case, the subject carries the highest degree of CD and in consequence conveys the high point of the message (acts as the ‘operative word/element’). In the latter case, another constituent than the subject takes over as conveyer of the high point of the message. This entails different relations between the constituents in regard to the dynamics of the communication expressed. Whereas in the former case the subject completes the development of the communication reflected by the sentence structure and nothing more is said about it in the sentence, in the latter case the communication offers some information about the subject.

All this entails different dynamic semantic functions (DSFs). In the former case, the subject performs the DSF of expressing a Phenomenon to be presented, animate or inanimate, concrete or abstract, real or imaginary. In the latter case, it performs the DSF of expressing a Bearer of quality. (Quality is to be understood here in a wide sense of the word: anything ascribed to the subject by the verb, or after a copula by an adjective or noun, is regarded as a quality, permanent or transient.)

If in Jan has come to the dining room the adverbial to the dining room is the only constituent that conveys retrievable information, the subject Jan performs the DSF of expressing the Phenomenon to be presented (the Ph-function), the notional component of the verb come performs the DSF of Presentation (the Pr-function). and the adverbial to the dining room performs the DSF of expressing a Setting (the Set-function).

If in Jan has come to the dining room the subject Jan is the only constituent that conveys retrievable information, it performs the DSF of expressing the Bearer of quality (the B-function), the notional component of the verb come performs the DSF of expressing a Quality (Q-function) and the adverbial to the dining room performs the DSF of expressing a Specification (the Sp-function).

The DSFs determine the thematic and the non-thematic (transitional and rhematic) sections of the sentence. The theme is constituted by (i) a context-dependent or context independent Bearer of Quality, (ii) and/or a context-dependent or context-independent Setting, (iii) and/or any other constituent that has become context-dependent. The rest of the sentence is rhematic.

‘Context-dependent’ can be defined the narrow sense of what is retrievable from the immediately relevant context and is in this way dependent on it. The sentence becomes embedded in context through the retrievability from the immediately relevant context.

What happens if in Jan has come to the dining room irretrievable information is conveyed only by the temporal exponent (for instance, has  may serve in a contrast to will: A: “Jan will come to the dining room” versus B’s correction “Oh, no, Jan has come to the dining room”), or by the temporal and modal exponents, which simultaneously provide a link and a boundary between the theme and the non-theme? The link between thematic and non-thematic information reflects a piece of information that is always irretrievable; (cf. Daneš, “Functional sentence perspective and the organization of the text” in Daneš. ed., Papers on Functional Sentence Perspective, p.90.)

The correction conveys a lot of retrievable (context-dependent) information. This is expressed not only by the context-dependent subject Jan performing the thematic B-function and the context-dependent adverbial to the dining room, performing the thematic Set-function, but also by the context-dependent notional component of come. Context-dependence has reduced the dynamic semantic status of this verbal notional component to that of a Setting, which of course is thematic. Under these circumstances, the high point of the message is conveyed by the temporal indication. Has performs the non-thematic Q-function. In this respect. it actually serves as rheme proper within the non-thematic section. As I have demonstrated in Firbas 1992, FSP in Written and Spoken Communication (pp. 88-93), the categorial exponents of the verb are multifunctional in FSP. Through its temporal and modal exponents, the verb form has come serves as transition proper.

The dynamic semantic functions are conceived of at the same level of abstraction as the syntactic functions. In the act of communication, a syntactic constituent performs a dynamic semantic function. A dynamic semantic function enables it to participate in the development of the communication in a co-determination of FSP. In this way the dynamic semantic function becomes one of the essential concepts of functional syntax, which accounts for the operation of syntactic constituents in the development of the communication. The dynamic semantic functions constitute the properties acquired by the syntactic constituents in the dynamics of the communication. They co-establish the thematic or non-thematic status of the syntactic constituents and participate in the conveyance of the communicative purpose.

Here I would recall one important conclusion drawn from the enquiry into the distribution of degrees of CD. It concerns the operation of the verb. The verb, or rather its notional component, shows a strong tendency to mediate between constituents carrying lower degrees of CD on the one hand, and higher degrees of CD on the other. This is manifested by the fact that the verb cannot complete the development of the communication in the presence of what has been termed a successful competitors: a constituent that owing to the interplay of FSP factors exceeds the verb in CD and prevents it from conveying the high point of the message. Through its mediatory function the verb, or rather its notional component, begins to build up the core (non-theme) of the message upon the foundation (theme) provided by the carriers of lower degrees of CD. This mediatory function is qualified as a transitional function between the theme and the rheme. (The rheme forms that part of the non-theme that is not taken up by the transition.)

It is important to add that whereas the notional verbal component shows a strong tendency to perform the transitional function, the categorial exponents of the verb, especially the temporal and modal exponents do so invariably. In this way, they perform the function of transition proper, and simultaneously provide a link and a boundary between the theme and the non-theme. (Like an ordinary transition, transition proper starts building up the core of the message within the non-theme.)

An important note must be added here. By the place of the ‘link’ or ‘boundary’ between the theme and the non-theme I mean its position in the interpretative arrangement. In contrast with the actual linear arrangement, the interpretative arrangement is the arrangement of elements carried out by the interpreter in accordance with a gradual rise in CD. (Whereas a linguist acquainted with the FSP theory does so consciously, a language user unacquainted with the theory does so subconsciously.) The actual linear arrangement is the real arrangement as it occurs on paper or the real arrangement in which the elements are perceived by the hearer.

It is important that emphasise of the FSP factors operates independently of the other factors participating in the interplay. The same naturally applies to the signals yielded by the factors. None of the signals operates independently of the other signals reflecting the interplay among FSP factors. So one and the same semantic and syntactic (grammatical) sentence structure can be applied under different contextual conditions. The contextual factor, which plays the dominant role, determines the extent to which the other factors can assert themselves.

Another important note concerns the difference between the intrinsic meaning of a linguistic element and its meaning modified for the attainment of the communicative purpose. The intrinsic meaning determines the semantic character of the element and predisposes the element to behave in a particular way in regard to the development of the communication. Yet a word, phrase, or any linguistic element conveying some meaning has its semantic content more or less modified in the course of the production of the sentence. Its semantic content is affected by the semantic content of the other elements of the sentence.

Significant modification thus takes place through the interplay among FSP factors. The semantic content is to adjusted to the communicative purpose to be fulfilled by the sentence. If in Jan has come to the dining room only the adverbial to the dining room conveys context-dependent (retrievable) information, come can assert itself as explicitly expressing appearance on the scene. In the presence of the context-independent subject, Jan, it recedes to the background and effectively participates in perspectivising the sentence structure toward Jan, presenting the person as a phenomenon appearing on the scene.

Under different contextual conditions, the modification results in different dynamic functions. (The qualification ‘dynamic’ indicates that the function has been acquired in the dynamics of the communication.) If in Jan has come to the dining room, only the subject Jan is context-dependent, the sentence structure cannot be perspectivised toward it. It is perspectivised away from it. The semantic feature of appearance/existence on the scene loses the power of participating in presenting Jan as expressing a person to be presented on the scene and consequently as conveying the high point of the message. This time the sentence is perspectived towards the adverbial to the dining room. It is important to note under the circumstances another semantic feature of come can assert itself. Together with to the dining room, come expresses motion and its goal. It has been demonstrated that as long as the goal of the motion is context independent information, it contributes more to the development of the communication and carries a higher degree of CD than the information of motion. It follows that both appearance/existence on the scene and the motion towards a goal are important semantic features that under certain types of interplay serve as important semantic dynamic signals. They do not, however, operate on their own. They can assert themselves as indicated only if permitted to do so by the other FSP factors. That they cannot operate on their own points to the systemic character of FSP.

In spoken language, the FSP factors are joined by the further factor of intonation. The signals it yields are prosodic features of different prominence: absence of stress, unaccented stress, accented stress, nuclear stress. (This gamut could be expanded.) These features are distributed over the sentence constituents, the sentence serving as a distributional field of prosodic prominence (PP). Intonation too does not operate on its own irrespective of the other factors and their signals. Intonation proves to be an effective modifier. It provides a running attitudinal commentary on what is conveyed by the constituents of the sentence. Conveying the speaker’s attitudes, intonation communicates information sui generis. In this way it participates in the development of the communication and is capable of raising degrees of CD. No rise in CD takes place if the distribution of CD as determined by the interplay of the non-prosodic FSP factors is reflected by the distribution of PP. Rise in CD is achieved through deviations from perfect correspondence between the two distributions. These deviations are functional and cause prosodic intensification of one constituent at the expense of another constituent or other constituents.

It is important to note that the deviations operate against the background of perfect correspondence between the two distributions. A comparatively rare, but highly emotive and effective, is a deviation which results in ‘re-perspectivising’ the distribution of CD as determined by the interplay of the non-prosodic factors. The effective character of this deviation is due to the very fact that perfect correspondence between the two distributions constitutes a background against which the deviations are evaluated. The perfect correspondence between the two distributions occupies a central position in the operation of FSP in spoken language.

 

2. On the status of the sentence

 

Is the sentence a basic unit? By interrelating the constituents, syntactic structuration delimits stretches of text as functional entities. The interrelations resulting in delimitation are rooted in semantics and signalled by formal features (through form). What has been delimited is a semantic and syntactic (grammatical) sentence structure capable of serving a communicative purpose, and of appearing in a functional perspective. In regard to form, Trost has come to the conclusion that a word or phrase that is syntactically unrelated both to what precedes and to what follows serves as a sentence. The verbal sentence is, of course, the fundamental form. Against it verbless structures serving as sentences can be assessed and evaluated.

 

3. On the meaning of ‘functional’

 

‘Functional’ has always suggested to me an operation performed by a linguistic element of any rank at the moment it is uttered and/or perceived. At that moment it serves, or participates in serving, to fulfil a communicative purpose of the language user. A ‘functional’ linguist is concerned with the way the task of the element is implemented. Seen in this light, the emphasis is on the operation aimed at fulfilling a communicative purpose imposed upon the element(s).

The terms in Czech aktuální, Russian aktual’nyi, German aktuel, and French actuel all pertain to the immediately relevant communicative step to be taken at to the moment of utterance and/or perception. Regrettably, English actual is a false friend; its primary meaning is ‘real’. The emphasis is should be on the hic et nunc situation happening now in the flow of communication. What is meant is the current operation of a linguistic element at the moment of utterance and/or perception. What links functional and aktuální is, of course, the communicative purpose to be fulfilled.

In regard to the development of the communication, the process of fulfilling the communicative purpose can be felicitously viewed as something ‘dynamic’ and ‘ongoing’.

 

4. Predictability vs retrievability

 

The development of the communication continuously narrows the context relevant to the immediately communicative step to be taken. In this way the immediately relevant context is created. There is a difference between a retrievable piece of information and a merely predictable one. Retrievable information is actually expressed and in this way ‘physically’, ‘tangibly’ present in the immediately relevant context text and therefore retrievable in the true sense of the word. The actual presence of a piece of information and the possibility of its re-expression qualify the information to act as a signal yielded by the contextual factor provided it has not been re-expressed beyond its retrievability span. This does not hold for predictable information, which has not been expressed yet.

The signal of retrievability simultaneously signals thematicity, even though  through the interplay of factors thematicity can also be signalled by irretrievable information; thematicity is not invariably linked with retrievability Still, an element conveying retrievable information cannot be rhematic. It could, however, be induced to convey some additional irretrievable information that dominates over the retrievable information. This would happen, for instance, if the pronouns you and me became bearers of contrast in He did not mean you, but me. On other types of predominant additional irretrievable information, see my paper “Retrievability span in functional sentence perspective”, Brno Studies in English 21/1995, pp. 17-45, pp. 21-23. A rhematic element always conveys information that is fully irretrievable or that dominates over retrievable information simultaneously conveyed by the element.

An important notion here is that of the retrievability span: the stretch of text in the course of which a piece of information remains retrievable. A piece of information becomes retrievable the moment it enters the flow of communication. In regard to a particular act of utterance and/or perception , however, it gradually loses its retrievability, if it is not re-expressed. An analysis of co-referential strings (constituted by elements sharing the same referent) has indirectly led to the conclusion that the retrievability span ends in the region of 6 through 8 sentences after the last expression of a given piece of information. The immediately relevant context is then constituted by all the retrievability spans that are open (live) at the moment a sentence is produced and/or perceived. The region of 6 through 8 sentences creates a borderline area between the immediately relevant context and the rest of the context. Within this area cases of potentiality may occur permitting different interpretations of retrievability/irretrievability. For a detailed discussion of the delimitation of the retrievability span and the delimitation of the immediately relevant context, see Firbas, “Retrievability span” in Brno Studies in English 21/1995.

The different arrangements may not sound equally natural because some of them may not conform to the predominating patterns of collocation and colligation. Nevertheless, the different arrangements can produce different functional perspectives.

 

5. On ‘preparedness’

 

It is certainly true that while producing a text the language user keeps his/her mind on the aim of the text and prepares the addressee for what is to come. S/he may not do so with unvarying consistency, but the unfolding of a communication constantly requires preparation. When opening a text, written or spoken, s/he she narrows the contextual section of common knowledge shared with the addressee down to the particular sphere s/he wishes to communicate about.

In regard to the immediately relevant communicative step to be taken, this contextual section is further narrowed down radically in order to comply with the immediately relevant communicative purposes to be fulfilled. In this way. the immediately relevant verbal and situational context is created. What is retrievable from the wider context of common knowledge existing beyond the immediately relevant context need not necessarily be retrievable from the immediately relevant context. Notions that are not explicitly expressed in the immediately relevant written context or notions that do not have their referents in the immediately relevant situational context are irretrievable.

 

6. On Presupposition

 

It is difficult to determine retrievability and irretrievability on the grounds of presupposition. Presupposition does not go by the signals offered by the dominant FSP factor, which is the contextual factor: the actual presence of pieces of information in the immediately relevant written context and/or the actual presence of their referents in the immediately relevant situational context. It operates with unexpressed assumptions that are in the interpreter’s mind. Owing to the interplay of FSP factors words expressing these assumptions may either be thematic or non-thematic.

 

 7. On Articles

 

The definite article cannot be invariably linked with retrievablity, and hence with thematicity. May I recall that full retrievability or predominant retrievability of an element is a signal of thematicity, but the thematicity of en element can be cosignalled by irretrievability.

 Likewise, the indefinite article cannot be invariably linked with rhematicity. Not operating on their own, independently of other signals, the definite and the indefinite articles can only signal thematicity and rhematicity, respectively, in co-operation with other signals and through the interplay with them. True enough, they respectively indicate known and unknown information in regard to the context shared by the sender and the addressee, and in consequence, information retrievable or irretrievable from this wider section of context. But in regard to the immediately relevant communicative step to be taken, known as well as unknown information is assessed in regard to retrievability and irretrievability, respectively, conceived of in the narrow sense, i.e. in regard to the immediately relevant context.

Still, under favourable conditions, the interplay of signals can induce the definite and the indefinite articles to effectively operate throughout stretches of text and respectively co-signal thematicity and rhematicity in such a way as to participate in signalling the thematic and the non-thematic (transitional and rhematic) layers through these stretches. (cf. my analyses of texts in Firbas, “On the thematic and the rhematic layers of a text”, in B. Warvik, S.-K. Tanskanen and R. Hiltunen, eds., Organization in discourse, Turku 1995.)

 

8.  On the function of the subject in regard to retrievability/irretrievability

 

(i) A young king ruled his country despotically.

(ii) The young king ruled his country despotically.

(iii). He ruled his country despotically

(iv) He did so.

 

The sentence structure (i) can open a narrative. Like the rest of the sentence, the subject A young king is context-independent. The structure is perspectivised toward despotically performing the dynamic semantic function of Further specification and acting as rheme proper.

The most natural contextual application of (ii) would render the subject context-dependent.  If context-dependent, the dynamic semantic status of his country would be reduced to a Setting; if context-independent it would serve as a Specification. As for despotically, it would serve as a Specification in the former case and as a Further specification in the latter. In either case it would act as rheme proper.

A marked application of (ii) could open the narrative. Markedness is evoked by the use of the definite article. The information conveyed by the subject is irretrievable not only from the immediately relevant context, but also from the rest of the context. Under the circumstances the definite article merely creates the impression of retrievability. It does not signal retrievability, but merely presents the information as if it were retrievable. This opening of a narration is in fact a stylistic device. The reader or listener is very well aware of the absence of a genuine retrievability signal. With due alterations, the same applies to the applications of (iii). The marked effect is enhanced by the use of the pronoun He.

In (iv), the high point of the message (rheme proper) can be conveyed by did. It would act as a pro-form stating that the activity in question actually took place and would carry the semantic feature of assertion. In the absence of elements surpassing it in CD — in the absence of successful competitors — did contributes most to the further development of the communication and carries the highest degree of CD. On account of its temporal and modal exponents, it simultaneously performs the function of transition proper. (On FSP and the multifunctionality of the verbal categorial exponents, see Firbas 1992, pp. 88-93.

To be sure, I have not discussed all contextual conditions, covering all the contextual applications or the entire contextual applicability of the structures adduced. Any element of a sentence structure can operate in the theme or the non-theme. My concern was to show that a subject not conveying the high point of the message and therefore performing the thematic B-function expresses either context-dependent (retrievable) or context-independent (irretrievable) information.

 

9. Comments on some of Beaugrande’s example sentences

 

(20) we had the letter typed at the dacha; some helpful assistant saw to it