If ThatAuthor(s): Deirdre WilsonSource: Linguistic Inquiry, Vol. 1, No. 3 (Jul., 1970), pp. 369-373Published by: The MIT PressStable URL: http://www.jstor.org/stable/4177579 .

Accessed: 14/06/2014 16:13

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact support@jstor.org.

.

The MIT Press is collaborating with JSTOR to digitize, preserve and extend access to Linguistic Inquiry.

http://www.jstor.org

This content downloaded from 62.122.77.48 on Sat, 14 Jun 2014 16:13:17 PMAll use subject to JSTOR Terms and Conditions

SQUIBS AND DISCUSSION

(i 6) This is a difficult book ' ] to bind to read to eat to copy to print to color to wrap to burn

IF THAT Deirdre Wilson, MIT

Certain grammatical sentences become unacceptable or odd on the addition of the words if that, while others do not. For example:

This log will burn for I 2 hours, if that. *This log will burn for a long time, if that.

500 people were there, if that.

r*Many *Several people were there, if that.

*ome *Those

The problem is to find a rationale for the behavior of if that.

I. There is an obvious difference between the starred and unstarred sentences above: the acceptable versions contain definite quantifiers (12, 500), while the starred versions contain indefinite quantifiers (a long time, many, several). A first suggestion might be that a sentence will take if that only if it contains a definite quantifier. However, in the following sentences, the acceptable versions are those containing indefinite quantifiers; the definite versions are unacceptable or odd:

This log will burn for a few minutes, if that. *This log will burn for 50 years, if that. A few nice people were at the party, if that.

*7,ooo nice people were at the party, if that. *At least 5o nice people were at the party, if that.

Moreover, if the quantifier becomes absolutely definite, the addition of if that is entirely ruled out:

*Exactly I 2 people were there, if that. *484 people stood up, if that.

Thus, it is the inaccuracy rather than the accuracy of the sentence in question which permits the addition of if that.

This content downloaded from 62.122.77.48 on Sat, 14 Jun 2014 16:13:17 PMAll use subject to JSTOR Terms and Conditions

370 SQUIBS AND DISCUSSION

In fact, the whole question of quantifiers is irrelevant. Consider the following:

*The fire was blazing, if that. The fire was smoldering, if that.

*I collided with him head on, if that. I grazed his fender slightly, if that.

Here neither version contains a quantifier. The difference between acceptable and odd forms seems to lie in the strength of the verb. One could claim of a fire that it was smoldering, flickering, burning, blazing, or raging, the first two claims containing weak versions of the verb to burn, the third neutral, and the last two strong. It is only the weakest two of these claims that can be modified by the addition of if that. There is a straightforward test for weakness in verbs of this type: those verbs that take if that will also take such words as hardly, barely, scarcely, only, not more than, etc. Those that are odd with hardly, etc. are also odd with if that:

*The fire was hardly blazing (nonsarcastic intonation). The fire was hardly smoldering.

*1 barely collided with him head on. I barely grazed his fender.

Hardly, barely, etc. function to weaken the claims made by the sentences in which they occur. They are plainly out of place with words which would not be chosen except for their strength: blazing, speeding, screaming, hurtling, and so on. Words such as these may always be replaced by a neutral or weaker word, if a weaker claim is intended. However, where a weak claim is already being made, and the speaker would like to suggest that even this may be exaggerated, hardly, barely, etc. are used, indicating that the claim being made is an absolute ceiling. If that appears to function in a very similar way. Thus, if that indicates that the claim being made by the sentence to which it is attached is an absolute ceiling: that the sentence in question is to be taken as possibly inaccurate, and as inaccurate, if at all, by way of overstating the case. If that is a natural minimizing device: "S, if that" might be paraphrased "S, or not even S"; "S, or perhaps not as much".

II. The problem of predicting the behavior of if that is largely that of predicting the behavior of hardly, barely, scarcely, only, etc. The oddness of certain sentences with if that is accounted for in the light of the above paraphrase.

Many, several, and some cannot cooccur with if that. They will not take hardly, etc.:

This content downloaded from 62.122.77.48 on Sat, 14 Jun 2014 16:13:17 PMAll use subject to JSTOR Terms and Conditions

SQUIBS AND DISCUSSION

*many

Hardly *soev people were there. someI

<*thoseJ

Their inaccuracy is not of the sort which the addition of if that could significantly modify: it is nonsense to talk of "many, or perhaps a few less", "several, or perhaps a few less", and so on. Moreover, they all make strong, or more than minimal, claims about numbers, and should not take if that for this reason. A few, on the other hand, does make a weak claim about numbers, and it is to be predicted that it will take if that. It should also take barely:

Q: How many nice people were there?

A: Barely (even) a few. A:A few, if that. f

Few is a special case. On the one hand, it makes a weak claim, and so should take if that; on the other hand, it is very similar in its behavior to many, which does not take if that. In fact, it takes neither barely nor if that:

*Barely few were there. *Few were there, if that.

However, it has its own form of minimization: one can say both "Few were there, if any", and "Barely any were there". There are a number of further minimizing devices of this type: for example, if ever, if at all, if anywhere, and if anything:

I saw him once or twice, if at all. I saw him rarely, if ever. He'll be in the bar, if anywhere. He was slightly drunk, if anything.

III. Rules for generation of these forms would involve two processes:

(i) Placement of a feature [+ Min] on the sentence to be qualified, plus some selectional restrictions on type of sentence (e.g. Time, Place, Manner, Number) which will take one or more of the forms: if ever, for example, would select a Time minimization.

(2) Characterization of constituents or sentences which take this feature. Hardly, barely, scarcely, only, not more than, etc., and weak verbs of the type mentioned in II, would directly carry the [+Min] feature, and any sentence in which they occurred would optionally take if that.

This would provide a partial syntactic characterization of the largely semantic facts mentioned in the analysis. It

This content downloaded from 62.122.77.48 on Sat, 14 Jun 2014 16:13:17 PMAll use subject to JSTOR Terms and Conditions

372 SQUIBS AND DISCUSSION

should be noticed, however, that in actual discourse, many sentences which do not actually contain hardly, etc., and which would normally be used to make neutral or even strong claims, may take a minimizing device. These sen- tences would be responses to even stronger claims, and would place an absolute ceiling only relative to those stronger claims:

A: Malraux is our greatest novelist, thinker and Minister of Culture.

B: He's your greatest Minister of Culture, if that.

To accommodate these sentences, the rules would have to be relaxed to allow the [+ Min] feature to be placed on any sentence which could take only, not more than, etc. Alternatively, since these sentences have a characteristic stress pattern, the feature might be placed in terms of this.

IV. It remains to inquire whether there are devices similar to if that, but which have the opposite function, which, where the accuracy of the antecedent is in question, serve to maximize, not minimize it, to qualify it as an absolute floor, rather than an absolute ceiling. Plainly, such devices exist:

500 people were there, if not more. *Many people were there, if not more.

He saw her frequently, if not more often.

There seem to be fewer of these forms, but this is an illusion: they are simply constructed in different ways. The same effect would be achieved by:

She was pale, if not pallid. She was unhappy, if not downright despairing.

Sentences that will take this construction in the maximizing sense will also take the addition of at least, or certainly:

She was pale at least, if not pallid. She was certainly unhappy, if not downright des- pairing.

However, these forms seem to be ambiguous in a way in which the minimizing forms are not. Consider, for example:

(i) He is one of the greatest minds of the century, if not of all time.

and (2) This is one of the best books of the month, if not

of the year.

Both (I) and (2) seem to be ambiguous as to whether they

This content downloaded from 62.122.77.48 on Sat, 14 Jun 2014 16:13:17 PMAll use subject to JSTOR Terms and Conditions

SQUIBS AND DISCUSSION

are claiming the truth, or conceding the falsehood, of the consequent. (i) might be paraphrased:

(a) He's not one of the greatest minds of all time, I agree, but he is one of the greatest minds of the century.

(b) He's certainly one of the greatest minds of the century, and probably of all time.

(2) is similarly ambiguous. Certainly, intonation would help to disambiguate the two, but it is an open question whether the deep structures are different. If not, this would be a further indication of the necessity of including aspects of something other than deep structure in semantic interpretation.

PERFORMATIVES AND TOKEN- REFLEXIVES

J. F. Staal, University of California at Berkeley

Disregarding several of the problems dealt with in Ross (in press), I shall in most of the following discussion simply assume that every declarative sentence S, is derived from an underlying structure of the form (I):

(I) SO

I say NP

Si

by applying a performative deletion transformation TPD such as (92) in Ross' paper.' Ross' proposal raises the ques- tion (which he treats only partially in 3.2. of his paper) of what happens to embedded declarative sentences. Ross has in fact argued that TPD can be applied only at the highest level. But if this transformation might also apply to an embedded structure of the form (i), the way is opened for certain ambiguities to arise in some cases, e.g. for the generation of a class of strings each with two different under- lying structures, viz. (2) and (3):

(2) so

I say NP

Si

S2

1 I am very grateful to Paul Grice, to David Perlmutter, and especially to Haj Ross for comments on an earlier draft of this paper.

This content downloaded from 62.122.77.48 on Sat, 14 Jun 2014 16:13:17 PMAll use subject to JSTOR Terms and Conditions