Sarkadi Zsuzsanna ANN-246/b

The status of whether: complementiser or WH-phrase?

English sentences consist of a combination of several layers constructed around functional and lexical heads. Looking at (1) it seems proved that declarative and interrogative clauses can be co-ordinated, thus, for only the members of the same category can do this, they must have the same outmost functional layer, the complementiser phrase. (1) He will drink but will he eat? Within ordinary main clauses, the CP layer does not always have to be filled in order to accept the clause as declarative. The case is the same with most embedded declarative clauses. (2) (3) [CP I know [CP [IP He likes water]]. [IP he likes water]].

Therefore when we want to express a question something is necessary to distinguish it from a declaration. In direct questions a WH-phrase standing for the missing information and (or only) the Subject-Auxiliary Inversion that moves the auxiliary positioned in I to the empty C position is sufficient for this. (4) (5) What does he like? Does he like water (or not)?

In embedded questions, however, it is important to know the type of the question reported. Variable questions like (4) are replaced with open interrogative clauses featuring WHphrases. At the same time, polar (and considering the possibility or not: alternative) questions like (5) are turned into those called closed interrogative clauses. (Huddleston & Pullum 2002: 870). (6) (7) I don’t know what he likes. I don’t know whether / if he likes water (or not).


after which the finite clause reports an information question requiring a statement as an answer. for whether. First of all. namely. introduces the embedded interrogative clause. thus it could not have been moved to the front from somewhere else. the case is not so simple. We could as well as stop at this point and just simply say that in order to replace variable questions a WH-phrase is moved into SpecCP and with polar or alternative questions the complementisers whether or if are used. at first sight it would seem obvious that they belong to the same category. Regarding the illocutionary force there is no problem. WH-phrases usually stand for a piece of information missing from the clause and get to the front by movement. At the same time. though. Consequently. Whether. however. for they can be analysed similarly. (Huddleston & Pullum 2002: 973). it must have been born in that position just like a complementiser. there are certain differences in their usage which make this correspondence doubtful. it cannot front a direct question like all other WH- 2 . However. Since it is in most cases interchangeable with if. whereas the non-finite clause reports a direction question requiring an order as an answer. In this essay. no information is missing.Sarkadi Zsuzsanna ANN-246/b Seemingly either the SpecCP or the C of the embedded clause has to be filled by an element that carries the feature [+WH]. Nevertheless. The dictionary defines whether as a conjunction used when we are talking about a choice between two different possibilities. – Here it has to be noted that WH-phrases are usually used as relative as well as interrogative pronouns. seemingly does not replace any part of the clause. we will not handle them separately. complementisers by definition – being heads of the clauses – should determine the properties of them. which means that it does not determine the finiteness of the clause. just like if. that they are both complementisers. the clause introduced by whether can be either finite or non-finite. However. In this respect it resembles the WH-phrases. For similar reasons.

the sentence remains grammatical. For comparison. only whether can be used when the interrogative clause precedes the predicate (Huddleston & Pullum 2002: 973). [ What he likes _ ] remains unclear. (10) [ That he likes water ] is normal. (8) (9) [ Whether / *if he likes water] remains unclear. Besides. Nevertheless. whereas whether does not act like this. while the use of WH-words – though is usually not the preferred variation – is not considered ungrammatical (Swan 2002: 456). It is obvious that the distribution of whether overlaps with that of if. (11) [ For him to like water ] is normal. sentences beginning with a declarative clause and complementiser are also grammatical. if cannot be used when the interrogative clause is the complement of a preposition while whether can (Huddleston & Pullum 2002: 974). (13) Tell me (about) what he likes. Thus the phenomenon described could also be explained by the deficiency of if and not of whether. when it follows the first co-ordinate. Many of the differences between them may be accounted for by the difference of category. In addition. the complementiser that can neither be used this way.Sarkadi Zsuzsanna ANN-246/b phrases. The only difference is that what evidently fills in the gap left by the missing information. 3 . So far we have stated that whether can take either a finite or a nonfinite clause as complement. However. if cannot be used when or not immediately follows the subordinator. only embedded clauses. (12) It depends on whether / *if he likes water. (14) I know about that he likes water. In this case it behaves just like WH-phrases do in pseudo-cleft sentences. In addition. but this does not mean that they are the same.

This explains I-to-C movement in direct questions: the auxiliary from I can move to the empty C and take the position behind the overt WH-phrase. and they are obviously not predicates. (18) [CP Whatc canb [IP hea tb [VP like tc]]]? The CP layer cannot be multiplied as many times as we want. the analysis of multiple questions provides further evidence that whether belongs to WH-phrases rather than to complementisers. (19) a. whether can be co-ordinated with not whereas if can only do this when the sentence is constructed differently. Moreover. For only the members of the same category and predicates can be co-ordinated. What does he like why? b. since WH-movement is local. when a direct multiple WH-question is turned to embedded. The trace occupies the SpecCP thus no other WH-phrase can try to move 4 . they can do this only step by step through all the SpecCP-s of the intervening CP-s while it leaves a trace at each stop. (16) I don’t know whether / if he likes water or not. The Doubly Filled COMP Filter defines that in a clause there can be either an overt WH-phrase in SpecCP or an overt complementiser in C (Haegeman 1994: 382). therefore in multiple WH-questions when more than one piece of information is asked for we cannot move all WHphrases to the front of the clause. for some reason. Seemingly. whether can be co-ordinated with WH-phrases while if cannot. only one of the WHphrases can be moved to the front of the sentence – and. (17) I don’t know whether / *if or why he likes water. As will presently become apparent. this evidence shows that whether should be considered to belong to the same class as why rather than if. * What why does he like? Similarly.Sarkadi Zsuzsanna ANN-246/b (15) I don’t know whether / *if or not he likes water.

*[CP_ _ [IP She will ask [CP whati _ [IP he knows [CP whether _ [IP he likes ti]]]]. As we have expected. c. just as if there were an abstract [+WH] element in SpecCP (Haegeman & Guéron 1999: 180). [CP _ _ [IP She will ask [CP whati _ [IP he knows [CP ti _ [IP he likes ti why]]]]. It appears that originally it had a meaning 5 . e. c. * [CP _ _ [IP She will ask [CP whati _ [IP he knows [CP ti if [IP he likes ti]]]]. surprisingly. However. that is a complementiser introducing declarative clauses which could stand without any complementiser at all. if we analyse whether as a WH-phrase (21.Sarkadi Zsuzsanna ANN-246/b even to the originally also empty intervening SpecCP (Haegeman & Guéron 1999: 177). if. admittedly. (21) a. In comparison. b. Finally it has to be mentioned that The Oxford English Dictionary (Simpson & Weiner (ed. Based on this evidence it has to be considered which analysis of whether is more probable. d. b.d). If whether is taken for a complementiser (21. the complementiser that or its null-variant do not block the WH-phrase moving towards the higher CP. [CP _ _ [IP She will ask [CP whati _ [IP he knows [CP ti _ [IP he likes ti]]]].e). we should try to create similar sentences with our complementisers and observe whether they are grammatical or not.* [CP _ _ [IP She will ask [CP whati _ [IP he knows [CP whyj _ [IP he likes ti tj]]]]. does block WH-movement. (20) a. On the basis of all this. [CP _ _ [IP She will ask [CP whati _ [IP he knows [CP ti that [IP he likes ti]]]].) 1989) still cites whether as “an interrogative particle introducing a disjunctive direct question”. it has to stand together with this incomprehensible [+WH] element which occurs with if whereas nothing similar can be found beside that – although. *[CP _ _ [IP She will ask [CP whati _ [IP he knows [CP ti whether[IP he likes ti]]]]. it seems quite understandable why it blocks WH-movement: it is sitting in SpecCP and as such it blocks the movement of any other WH-phrases. whereas it is obvious even from the dates next to the examples that it has not been used as such for a few hundred years now. [CP _ _ [IP He knows [CP whati _ [IP he likes ti why]]]].

(22) Does he like water or not? (23) Which of the two possibilities is true: that he likes water or that he does not? (24) I don’t know which of the two possibilities is true: (that) he likes water or (that) he does not. it inquires into the verity of a whole clause. Although it can be used in similar contexts as the complementiser if. 6 .Sarkadi Zsuzsanna ANN-246/b similar to the Latin interrogative pronoun uter (‘which of the two?’). in not all the cases – examples (8).c) – was it evident why if behaves so differently from its declarative complementiser counterpart that. (15). (16). and how a more complex sentence is constructed around them. Unlike other WH-phrases and more like the phenomenon of SAI. this would require further investigation. Its meaning is a little more complex that that of the other WH-words. (25) I don’t know whether he likes water or not. In conclusion. Over the centuries its meaning has altered and narrowed – although it still has the original disjunctive meaning – while it kept its position in the sentence. possibly it is derived from the same stem. However. Nevertheless. maybe this is why it ceased to be used to front direct questions – especially because the same meaning can be expressed with less effort using simply the SAI. (21. there are too many differences in their usage concerning what complements they can take. to what word classes they can become complements. it can be stated that there is more evidence in defence of the view that whether is a WH-word rather than a complementiser. Maybe this can be accounted for by the mere fact that it is interrogative and not declarative. This would provide a plausible answer for why whether does not stand for a piece of information missing from the sentence.

A. Haegeman. G.. L. Guéron.Sarkadi Zsuzsanna ANN-246/b References Haegeman. Oxford UK and Cambridge USA: Blackwell. (2nd ed. J. The Oxford English Dictionary XX. L. Swan. C. New York: Cambridge University Press. (1994). K. E. (2002).. Pullum. R.) Oxford: Oxford University Press. Huddleston.) Oxford: Clarendon Press. English Grammar: A Generative Perspective. Oxford: Blackwell. (1999). J. 7 . S. Introduction to the Government and Binding Theory. The Cambridge grammar of the English language. Weiner. Simpson. (editors) (1989).. (2nd ed. (2002). M. Practical English Usage. Cambridge.

Sign up to vote on this title
UsefulNot useful