4

The body was no longer twitching. The skin had taken on a milky bluish tinge. The corner of the mouth seemed to have stopped bleeding, and what little blood was still visible now appeared very slightly darker and thicker, although the red, green and amber bulbs of the footlights might be influencing my color perception.
[The Weed That Strings the Hangman's Bag]

There are two verbs after what-phrase. In this case, (1) is relative pronoun ‘that’ omitted after blood, or (2) does ‘what’ take two roles of a relative adjective and relative pronoun?

Robusto
  • 151,571
Listenever
  • 3,454
  • 7
  • 34
  • 70

3 Answers3

4

The best classification of what in this usage is as a determiner:

what: determiner

...

2. a. the (person, thing, persons, or things) that: we photographed what animals we could see

...

Collins English Dictionary – Complete and Unabridged © HarperCollins Publishers 1991, 1994, 1998, 2000, 2003

what little blood was still visible is a noun clause used here as subject; it may be paraphrased the small amount of blood that was still visible, or such traces of blood as were still visible. This what-structure does not take that.

Little is a quantifier here rather than the categorially polysemic adjective.

Edwin Ashworth
  • 83,272
  • In the OP's provided sentence, it'd be a relative determiner, yes? – coleopterist Jan 24 '13 at 13:50
  • 1
    Sorry about the delay - it's hard to find analyses. In we photographed those animals, those would be a demonstrative determiner (with a deictic - extra-textual pragmatic - role). In we photographed those animals that we could see and we photographed what animals we could see, I think you must be right - there is an adjective clause following, so the term relative applies. So, some but not all determiners have a syntactic role. – Edwin Ashworth Jan 24 '13 at 16:54
  • 1
    At http://en.wikipedia.org/wiki/English_determiners , the subclass is illustrated: Relative determiners: which (quite formal, as in He acquired two dogs and three cats, which animals were then...); also whichever and whatever (which are of the type that form clauses with no antecedent: I'll take whatever money they've got). Seems correct to me. And, at http://simple.wiktionary.org/wiki/Category:Relative_determiners : Pages in category "Relative determiners"

    The following 4 pages are in this category, out of 4 total. W what whatever which whichever

     – Edwin Ashworth Jan 24 '13 at 20:51
3

What is here an adjective, meaning in the OED’s, definition 8, that (or those) . . . which (or who); such . . . as; often expressing quantity, So much (or many) . . .as.

One of eight citations illustrating the use is this, from John Ruskin: ‘I will take what indulgence the . . . reader will give me’.

Barrie England
  • 140,205
3

Neither term is correct for what. What (like how) is a Wh-word that can't be used as a relative clause marker. For example, the following are all relative clauses marked with Wh-words; the last two are ungrammatical -- in practice, we use that instead of how to mark relatives.

  • the man who did it
  • the place where they did it
  • the day when they did it
  • the reason why they did it
  • the thing which they did
  • the way that they did it
  • *the thing what they did
  • *the way how they did it

As for the difference between "relative adjective" and "relative pronoun", pay no attention; relative markers can function as both. In fact, pay no attention to what "Part of Speech" a word is said to be; the citations are wrong, they're not using the correct list anyway, and, in the first place, most English words belong to several different categories.

However, since this is not a relative clause, this must be a different what. It is in fact the Wh-word that introduces an embedded question subject complement.

  • what little blood was still visible

which is a noun clause that is the subject of appeared in the quotation.

Embedded question complements are finite (they require a tensed verb like was), and they must begin with a Wh-word (there is in fact a special Wh-word whether, which is used only in embedded questions to mark an embedded Yes/No question).

  • Will she be here?
  • I don't know whether she will be here.
  • Whether she will be here is something no one knows.

This kind of embedded question refers to the answer to the question, whether it's known or not:

  • What they did was terrible. = The answer to "what did they do?" is terrible.
  • What they did is unknown. = The answer to "what did they do?" is unknown.

And, to get finally to the point, in this case it's a special use of what, quantifying the amount of blood and indicating its degree.

  • what little blood was visible = The answer to "what amount of blood was visible" is "little".

Other synonymous constructions that can be substituted in the same context include

  • the small amount of blood that was visible
  • the small amount of visible blood
  • what visible blood there was
  • what blood there was (that was) visible

Afterthought
I have since come to believe that there is no semantactic difference between "headless relative clauses" and "embedded question clauses".

Whatever distinction there may be becomes ever less pronounced the more the clauses are dismembered, especially when they reach skeletal infinitives like What to do or How to do it. I just call them all "Wh-clauses" any more.

John Lawler
  • 107,887
  • So, is relative determiner also inaccurate? Isn't "what little blood" referring to the (quantity of) bleeding earlier in the sentence? – coleopterist Jan 25 '13 at 07:24
  • Relative clauses are not involved in the question, so "relative" anything is inaccurate. If you're not referring to relative clauses, don't use the word relative; comparative is the term that refers to comparisons like this. – John Lawler Jan 25 '13 at 16:18
  • relative [ˈrɛlətɪv] adj
    1. (Linguistics / Grammar) Grammar denoting or belonging to a class of words that function as subordinating conjunctions in introducing relative clauses. In English, relative pronouns and determiners include who, which, and that Compare demonstrative [5] interrogative [3]
    2. (Linguistics / Grammar) Grammar denoting or relating to a clause (relative clause) that modifies a noun or pronoun occurring earlier in the sentence

    Collins English Dictionary – Complete and Unabridged © HarperCollins Publishers 1991, 1994, 1998, 2000, 2003

     – Edwin Ashworth Jan 25 '13 at 16:28
  • I think the adjectival nature of the those-structures is being referenced by the term relative here: We photographed those animals we could see. (determiner usage of those) We photographed animals - those we could see. (pronoun usage of those) – Edwin Ashworth Jan 25 '13 at 16:32
  • @EdwinAshworth: possibly that's the intention. But both definitions of relative that you cite refer to use in relative clauses. The fact that relative clauses usually modify a noun does not change the fact that it's relative clauses, not modification in general, that's being referred to. There are lots of non-relative constructions that modify nouns, too. – John Lawler Jan 25 '13 at 17:10
  • @John Lawler: So you disagree with the views (or at least the terminology) of the staff at 'Oxford Dictionaries'? (It's the sort of thing I often do): http://oxforddictionaries.com/definition/english/what what 2 [relative determiner] (referring to the whole of an amount) whatever: he had been robbed of what little money he had – Edwin Ashworth Jan 26 '13 at 16:43
  • This is a syntactic phenomenon, and dictionaries are not involved with grammar. They provide lexical information, and (if they're good) the actual grammatical categories that words belong to; most American dictionaries, for instance, don't give enough information to allow one to decide that *I ate a bowl of rices and *I ate a bowl of bean are both ungrammatical, but that they're both OK if one swaps the plural inflections. If you want grammar, you go to a grammar, not a dictionary. – John Lawler Jan 26 '13 at 17:12
  • 1
    We're still in the situation described by a Christian speaker I know: "Do you know why Bible teachers have so many different versions? So they can find one that says exactly what they want to say." Or even worse, since most grammars disagree at various points, we'll be left either confused or adopting a poor interpretation. Or unable to find a version we're satisfied with. – Edwin Ashworth Jan 26 '13 at 17:33
  • We don't have to solve the problems of the world like Christians do, and there is no Authority that censors grammatical terminology, so clearly you or anybody else can use any terms you like, for whatever reasons you like. Just don't expect that it's going to be understood by professionals, who have their own terminology and their own habits of usage for it. To a grammarian, unless specially qualified, relative in a technical grammatical term refers to relative clause. That's all, really. There are other senses of relative, but they're not grammatical. – John Lawler Jan 26 '13 at 17:44
  • Will http://www.docstoc.com/docs/122771064/POSTMODIFICATION-IN-NOUN-PHRASES do as a 'grammar source'? (P2) nominal relative clauses ... 7a What money I have is yours. (Then the oxforddictionaries.com definition given above follows.) – Edwin Ashworth Jan 26 '13 at 23:33
  • Not without authorship and citations, no. It's pretty good about the structure, but it does call these embedded questions "nominal relative clauses", whereas the syntactic literature calls them "embedded questions". And these in turn are confusable with "headless relatives" like Who steals my purse steals trash, where who comes from something like He who. – John Lawler Jan 27 '13 at 18:45
  • One of the biggest differences between embedded questions and relative clauses -- whatever they begin with -- is that Relative Clause Formation operates over variables, whereas embedded questions are cyclic and governed by the matrix verb. See Rule Typology. Though there are some circumstances where they merge and one can hardly tell the difference; syntax has a lot of bleed-offs like that between categories. – John Lawler Jan 27 '13 at 18:53
  • Note Afterthought above. – John Lawler Apr 17 '15 at 16:49