A friend of mine sent me a question from his nephew’s ninth-grade final English exam at Haishan High School in Banqiao, New Taipei City:
Which is the correct completion: (a) or (b)?
Lydia knows few things,
(a) and so does Peter.
(b) and neither does Peter.
Stare at that for a few moments and decide what your answer would be.
Here’s the puzzle: My friend discovered, by consulting various English speakers, that Americans all choose (a) as correct, while British and Australian speakers choose (b). This rather shocked me. Let me explain why.
The Cambridge Grammar of the English Language (CGEL) uses certain tests to diagnose polarity — the difference between positive clauses (like He denied it) and negative ones (like He didn’t deny it). Here are two useful reliable tests (CGEL, Chapter 9, §1.1; prefixed asterisks mark ungrammatical examples):
- The continuation not even is compatible with negative but not positive polarity. Compare He didn’t deny it, not even to journalists (but not *He denied it, not even to journalists). Note also We saw nothing, not even the entrance.
- Negative clauses take only positive confirmation tags (like is it?), not negative ones (like isn’t it?). Compare He didn’t deny it, did he? with *He didn’t deny it, didn’t he?. Note also This seems not implausible, doesn’t it?.
Notice that the presence of not or n’t won’t work as a test: That seems not implausible is positive (hence we get That seems not implausible, doesn’t it?).
The contrast between neither and so continuations does yield a test: Compare He didn’t deny it; neither did she with *He didn’t deny it; so did she.
Next I have to note that there are seven words called approximate negators (CGEL, Chapter 9, §3.3). These define negative clauses allowing a few exceptions. For example, rarely makes the clause They rarely enjoy it negative (notice it takes do they?, not don’t they? ). The determinative few is another approximate negator. Few reviewers liked the book has negative polarity (compare Few reviewers liked the book, did they? with *Few reviewers liked the book, didn’t they?).
We now face the question of whether few can negate a clause when it occurs in an object noun phrase. If so, then Lydia knows few things is negative, and should take a neither continuation, so the correct exam answer is (b); but if not, the sentence should take a so continuation, and the right answer is (a).
The trouble is that Lydia knows few things seems (at least to me) only slightly infected with negative polarity, but not completely. It’s on the cusp. I hate to think that a fundamental matter like negative vs. positive polarity might not be clear and sharp; but with approximate negators, that’s the way it seems to be.
Cross-checking with the other two tests does not fully remove the unclarity. Here are the sentences whose grammaticality we need to assess:
Lydia knows few things, doesn’t she? (should be ungrammatical)
Lydia knows few things, not even about her own country. (should be grammatical)
I’m not necessarily the best judge. My first-language acquisition was in England; about 40 percent of my adult life was in America; and I now work in Scotland among many American colleagues. My intuitive judgments on subtle points of transatlantic dialect divergence aren’t worth the paper they’re usually not written on. I’m unsure enough about the two sentences just cited that I would never base an exam question on them. But if forced to make a choice, I think I would lean toward saying that Lydia knows few things has negative polarity.
Haishan High School takes that view (though it generally uses American English teaching materials), and takes neither continuations to be a reliable test for negative polarity. Hence (b) is claimed to be the “right” answer. But my friend’s survey suggests that native speakers of American English will not get a perfect score on Haishan’s ninth-grade English exam.
In general, I don’t like exam questions of this sort as a way of testing non-native speakers. Too subtle, and not effective in diagnosing practical conversational command of English.
But the theoretical puzzle for me as a linguist is this: If my friend is right about the dialect split, how on earth could it be that Americans come to feel Lydia knows few things (and indefinitely many other such clauses) have positive polarity, while British and Australian speakers grow up assuming such clauses are negative? From what feature of ordinary everyday English discourse could they possibly be picking this up?
