Discourse functions and language-specific conditions for the use of cleft(-like) sentences: a prelude(1).
Ekkehard Konig (1991: 5)
Lambrecht (this issue) tells us that it was Jespersen who coined the term "cleft sentence" for a special case of relative clause that does not share the defining properties of other relative clauses. And it was Jespersen who pointed out the focusing functions of cleft and pseudocleft sentences and explained the typological variation across languages like English, Danish, French, and Italian as a compensatory device of languages with a comparatively rigid grammatical word order. The research program that has unfolded from this has brought ever-finer distinctions in the discourse functions of clefts. (For a survey of the most widely discussed views of the last twenty-five years or so see Mats Johansson, this issue.) To the dismay of scholars who aim at a more rigorous definition of linguistic phenomena, the term "cleft sentence" has been extended to related structures that cannot be "declefted" without a more or less extensive form of lexical adaptation (an extreme case of which is discussed in Doherty, this issue; a number of cases requiring minor adaptations are listed in Lambrecht, this issue).
Irrespective of whether we take a wider or a narrower view of cleft structures, comparative studies of linguistic phenomena will always raise the question of whether apparently analogous structures are really comparable. If we want to reach the goals set by the quotation above, we need a clear understanding of the theoretical concepts involved in the comparison. There are two major problems we have to cope with. All the concepts needed to determine cleft structures and their uses ("presupposition," "focus," "givenness," but also "relative clause," "basic order," "subject," "expletive," "monoclausal," etc.) are theory-dependent and cannot be exchanged freely between the various approaches. Thus, the discussion of the formal and functional properties of cleft structures may require at least some degree of familiarity with different linguistic frameworks (like functional or generative grammars, text-linguistic or discourse-representational theories, compositional or constructional approaches, and the like). This may be difficult enough, but the second type of problem seems to be even more difficult.
Even if we know what we mean when we classify an element as a focusing device in one language, how can we know that a superficially similar element in another language can be classified in the same way? The two elements participate in different networks of focusing and backgrounding devices, and if we want to compare the formal and functional properties of clefts across languages, we have to include the other language-specific devices with similar functions. Thus, certain cases of topicalization can be classified as focusing means, which places them in a line with clefts and other focusing devices. But while German, for example, allows topicalization of almost any element, "topicalization" in English is heavily constrained. (The similarity of position at the beginning of sentences is apparently relativized by the dissimilarity of the subsequent elements: finite verb in German, subject in English.) We can expect the English constraint on topicalization to be at least partly compensated for by clefts.
But how can we prove this? The papers contributed to this issue demonstrate that one can either make use of systematic variation of generated and attested examples (Lambrecht, Delahunty, Doherty) or collect a statistically significant amount of data from existing records (Mats Johansson, Stig Johansson). Either method relies on the researcher's intuition about the theoretically relevant elements of the findings and in particular on a clear understanding of the crucial concept of focus.
In my view, there is an aspect of focus that accounts for a great deal of the existing theoretical unclarity and contradiction: there is a linguistic side of focus, pertaining to prosody, syntax, and semantics, and there is a discursive side of focus, concerning progress in discourse. We can expect these sides to meet, in line with the presumption of optimal relevance. But if inappropriate linguistic devices have been chosen, the two sides will fall apart. If we can separate them theoretically, we can also obtain a better understanding of the specific properties of linguistic focus across different languages.
To come to terms with both sides theoretically, we need a theory of focus in discourse in addition to the prosodic, syntactic, and semantic theories of focus.
The statistically significant evidence from translations can serve as a guide to the theoretically relevant data. But it can serve as a basis for theoretically reliable conclusions only if we succeed in isolating the individual factors that contribute to a particular interpretation.
The same features may be marked more than once in an expression; or we could do without marking altogether within a particular discourse. The theoretically revealing cases are those in which a particular linguistic form is discourse-(in)appropriate as opposed to an alternative form within the same language or across languages. (The method of control paraphrases used in the comparison of original and translated sentences in discourse offers a systematic path to such data; cf. Doherty, this issue.)
As Delahunty shows, there are many ways to interpret the discourse relations in which an example participates, and it is most crucial for the theoretical interpretation of a variety of data to abstract away from the pragmatic influence of the individual case. This requirement concerns coherence relations in discourse and rhetorical relations.
Parallelism and contrast, the two major rhetorical relations, are the keys to focus interpretation in cleft sentences. They help us to separate the background material from the discursive focus or foci. Let me illustrate this by using an example. Consider the sequences of sentences from the famous dialogue between Humpty Dumpty and Alice.
Humpty Dumpty says,
(1) I shouldn't know you again if we did meet. You are so exactly like other people.
The first two sentences form a discourse segment that is related by the (coherence) relation of explanation. But if we attach Alice's answer,
(2) The face is what one goes by, generally.
to this segment, the three sentences together form a larger discourse segment with a common theme, saying something like "there is e one can go by (to recognize someone, as for example Alice)." The variable varies over the polarity of negative and positive specifications. Alice's answer specifies [Alpha] as "face" (and hence positive), while the preceding segment specifies the variable as negative, implying that there is nothing one could recognize Alice by. This means that the discourse relation between Alice's and Humpty Dumpty's statements is contrastive and requires a contrastive-focus interpretation of Alice's answer.
So much for the discursive side of focus in this example. What do we know about the linguistic side of focus, that is, about the formal means associated with focus interpretation? We know that a basic, declefted version of Alice's answer could not express the discursive focus on its own. The sentence
(3) One goes by the face, generally.
would be associated with a mere presentational focus per default, which may be a projecting, "wide" focus, presenting the entire proposition as new information. The default interpretation will have to be reanalyzed contextually, which requires extra effort as the discourse relation comprises more than just the preceding sentence. We can avoid the reanalysis if we make use of a formal clue marking the contrastive focus. In speaking, we could mark the focus exponent by strong stress, signaling a contrastive interpretation, but in writing we have to resort to other means to signal the discursive focus.
This is where cleft sentences come in handy, as their double-clause structure offers not only a split-focus domain with an existential presupposition but, in the case of the inverted pseudocleft, a formal clue to contrastive focus. While the it cleft signals a discourse relation to a preceding segment extending beyond the immediate context, it does not, on its own, signal a contrastive relation.
(4) It is the face one goes by, generally.
expresses the existential presupposition that there is [Alpha] that one goes by and specifies [Alpha] by the (presentational) focus in the matrix clause.
Anything else is left to the contextual interpretation. We are free to reanalyze the presentational focus of the matrix clause as contrastive (or to assign an additional focus to the embedded clause, which would be the case in an informative-presupposition cleft).
Although the it cleft is easier to process than the basic version, it cannot compete with the original formulation, the inverted wh cleft.
Without its inversion, the wh cleft would, however, be even less appropriate than the it cleft, as it blocks the contrastive reinterpretation of the wh clause. The sentence
(5) What one goes by, generally, is the face.
presupposes that there is something that one goes by and does not allow this to be negated -- which is, however, necessary, as we have to form a common theme subsuming Humpty Dumpty's view.
The inverted wh clause
(2) The face is what one goes by, generally.
meets all the discursive requirements by topicalizing the noun phrase in the copula clause. This type of cleft marks the "face" as contrastively focused. And although the subclause signals a common theme, it expresses the same presupposition as the it cleft, that is, it does not fix the polarity of the clause to the positive value.
All clefts indicate larger discourse relations, but the way they are integrated into the discourse differs according to their formal characteristics. All clefts pertaining to the canonical order signal a macrostructural relation of the parallel type, that is, a noncontrastive focus. All clefts signal a partitioning of the information structure, but only "canonical" wh clauses fix the background once and for all, and only the inverted wh clause marks its topicalized element as a contrastive focus. Everything else will be interpreted by default and, if need be, reinterpreted discursively. Clefts are syntactic discourse relators just as but and also are lexical discourse relators. We are, as authors, free to make use of them or to rely on our interpreter's capacity to figure out the discourse relations without any such clues.
If we compare clefts with other information-structuring devices, or their uses across languages, things become yet more complicated. Even if there are superficially similar devices available, as for example clefts in German, and even if they could be assumed to have similar information-structuring functions, they may prove inappropriate for the same discourse. Thus, the published translation of Alice's answer in German is
(6) Im allgemeinen richter man sich da nach dem Gesicht.
and not an inverted wh clause:
(7) Das Gesicht ist es, wonach man sich (da) im allgemeinen richtet.
The German cleft does not provide us with an easy-to-process sentence structure (cf. Doherty, this issue, on the processing difficulties of German copular structures), or an easy-to-process focus structure. But the cleft is not necessary anyway because we can associate the published version with an appropriate interpretation directly. The basic version of the German sentence is
(8) Man richtet sich im allgemeinen nach dem Gesicht.
Gesicht would be associated, per default, with a presentational focus, which could project onto the entire proposition (just as was the case with the English basic version). If we topicalize the adverbial,
(6) Im allgemeinen richtet man sich nach dem Gesicht.
we make use of an additional focusing device, also marking the adverbial as focused, albeit only as a presentational focus, expressing a topic shift within a parallel relation. This amounts to saying that the general procedure can be applied to the individual case under discussion, too, which is what Alice meant.
If we translate the German version back into English,
(9) Generally, one goes by the face.
the effect of topicalizing the adverbial is different. It marks a contrastive relation, which amounts to saying that the general procedure does not apply to the individual case under discussion, which is clearly discourse-inappropriate.
But why should the topicalization be associated with a contrastive interpretation in English and with a parallel interpretation in German? The answer lies in the different sentence structures determining the input and output conditions of the topicalized item. The basic position of the English adverbial is more distant from the landing position than that of the German adverbial, and -- in relation to the position of the finite verb -- the topic position of the English adverbial is "higher" than that of the German topic. If this analysis is correct -- which the analyses of many more examples suggest -- it shows that we need both a discursive-focus theory and a linguistic-focus theory to interpret the cross-linguistic findings.
Humboldt University, Berlin
(1.) Correspondence address: Institut fur Anglistik und Amerikanistik, Ubersetzungs-wissenschaft, Humboldt-Universitat zu Berlin, Unter den Linden 6, D10099, Berlin, Germany. E-mail: Monika.Doherty@rz.hu-berlin.de.
Konig, Ekkehard (1991). The Meaning of Focus Particles. A Comparative Perspective. London and New York: Routledge.
|Printer friendly Cite/link Email Feedback|
|Publication:||Linguistics: an interdisciplinary journal of the language sciences|
|Date:||May 1, 2001|
|Next Article:||A framework for the analysis of cleft constructions(*).|