Petra Hendriks. 2004. Either, both and neither in coordinate structures. To appear in: A. ter Meulen & W. Abraham (eds.), The Composition of Meaning. John Benjamins, Amsterdam.
bq. When the elements either, both and neither occur in a coordinate structure, they are usually analyzed as conjunctions. In this paper, it is argued that these elements are better analyzed as focus particles. The analysis of these so-called initial or correlative conjunctions as focus particles is motivated by their resemblance to focus particles with respect to (1) their distribution, (2) their interaction with sentential intonation, and (3) their contribution to the interpretation of the sentence.
Thank you Kai von Fintel for opening the debate.
It seems that one common and popular thing among the Dutch community is reclassification of words and phenomena. They like doing it. For instance, they take phrasal verbs and claim that the final word/morpheme should not be considered a preposition, but rather a particle. This paper is something in the same line: it argues that ‘either’ and ‘both’ should be classified not as conjunctions, but rather as focus particles.
Well, assuming that Hendriks’ claim is true, the question is what do we do with that. What this reclassification might mean? That we replace one empirical phenomena by another?
Or that this has more important theoretic consequences? I would like to hear the other readers on this question.
But I would be careful to accept her proposal so soon. In Portuguese there are at least two possible ways of making a contrastive focus. One way is by stressing the focused element. The second way is by converting the word order. Now, in the examples bellow the subject is focused in (B) but not in (A) (with neutral intonation):
(A) Tanto o Gil quanto ela fizeram bobagem.
(B) Fizeram bobagem [tanto o Gil quanto ela]focus.
‘Both Gil and her made something stupid’.
My intuitions about English is that one may focus a ‘both x and y’ argument, but it is not necessary that we always have focus when using ‘both…and…’. So I ask whether the native speakers agree to these intuitions of mine or not.
Anyway, there is one thing about Hendriks’ paper that is interesting from the cultural point of view. Her paper has examples only in English and she is Dutch. Well, for me there is no problem about it because I myself have written works with only English examples.
But when someone, who is not Dutch and is not an English native speaker, writes a paper using only English examples, the Dutch community reacts making comments like:
‘why don’t you give us data from your own native Language?’, ‘the best contribution foreigners may give is information about their own languages, we are interested in knowing about your own language’, ‘the idea of collaborating with foreigners is precisely to know about their own languages’, etc.
So, if one applies the Dutch rules to Dutch authors themselves, one should ask mevrouw Hendriks why hasn’t she written about Dutch. But, as I am not Dutch, I shall not exact it from her. Rather I reserve to myself and to other linguists the right to write about English or about anyother language, should it be his own native language or a foreign language.
May 19th, 2004, at 11:04 pm #As it seems that Hendrik’s work has not yet caused reactions from other readers, I take the liberty to add other comments:
Firstly I should ask which framework she is assuming. She quotes authors that approach the same issue from several perspectives, which is a sign of theoretic non-prejudice. But she should say which theory she is using. Is it GB? Minimalism? The anti-symmetric programme?
It is not possible to make an approach that is inter-theoretic or theory neutral.
Secondly, in pursuance of the previous comment, she should define the main concepts she uses to make her tests.
May 21st, 2004, at 4:28 am #For instance, she uses the notion of c-command several times, but in the literature there has been more than one precise definition of c-command. Does she use the c-command notion proposed by Kayne? Is it GB c-command? The difference is huge. A kaynean c-command notion excludes the possibility of a middle projection X’ reciprocally c-commanding a specifier.