No counting

Dec 19, 2010 by

In the previous posting I’ve mentioned that if we develop the prescriptivist idea that “two negatives in English destroy one another, or equivalent to an affirmative”, we would expect that three negatives would amount to a negative, four negatives — to an affirmative and so forth. However, this is not how it works. Not for the speakers who use multiple negatives (either number of negatives seems equally okay if you speak a language/dialect that allows multiple negatives in principle) and not for the prescriptivists who ban them in “standard” English. The reason that any number of negatives is equally good/bad is that language/grammar does not and cannot count.

In fact, this “no counting!” generalization is a corolary of the Structure Dependence Principle, which states that all rules of implicit grammars make reference to terms defined structurally, not in terms of linear order. For example, consider what rules describes the formation of yes/no questions in English. It must be something along the lines of “move the auxiliary in front of a preceding noun phrase that functions as its subject”. This rule makes reference to such structurally defined terms as “auxiliary” (non-last verb in a clausal verb sequence), “noun phrase” (structural unit headed by a noun) and “subject” (a notion that probably must be defined structurally too, but that’s a larger issue I will leave aside now).

Taking a declarative John has eaten a pie, this rule generates Has John eaten a pie?… however, if the element preceding the auxiliary is not a noun phrase subject (e.g., Down will come taxes) or the only verb in the sentence is not an auxiliary (i.e., there only one verb, as in John received a prize), the inversion rule doesn’t go through, hence the ungrammaticality of *Will down come taxes? and *Received John a prize?

However, the rule as formulated above is not sufficient. What if there are two auxiliaries in the sentence? Which is to undergo inversion? Take a sentence Australians [who are sunburned] are friendly. There are two auxiliaries are here. One could imagine two rules:

Rule A: Move the first occurrence of the auxiliary to the front.
Rule B: Move the auxiliary from the main clause (bracketed above) to the front.

Which is the rule of English? It must be the rule B, as the rule A generates an ungrammatical sentence:

Rule A: *Are Australians [who __ sunburned] are friendly?
Rule B: Are Australians [who are sunburned] __ friendly?

In fact, not only is the rule A not a real rule of English grammar, it is not a rule of any natural language grammar. No language forms questions (or any other constructions, for that matter) by operating on the first (or second, or third, and so on) occurence of the auxiliary (or of anything else). There are simply no rules that “count” in natural language. Nor do children ever postulate a rule like the rule A above, that is they never produce mistaken questions like *Are Australians [who __ sunburned] are friendly?… as has been verified empirically in a study by Crain & Nakayama (1987).

Moreover, Smith & Tsimpli (1995) conducted an ingenious study further proving that natural languages do not count. They tried to teach the polyglot savant Christopher (and a group of controls) an invented language Epun, which had rules that were structure-independent but relied on counting instead. For instance, emphatic sentences in this made-up language Epun were formed by adding a suffix to the third word of the non-emphatic sentence. Interestingly, neither Christopher nor the controls were able to deduce the rule, although they could easily solve a comparable ‘counting’ problem in a non-linguistic area.

But if language doesn’t count, how are rules that make reference to the second position, such as the verb-second construction in Germanic languages, possible? This is a topic we will return to in a future posting.

Related Posts

Subscribe For Updates

We would love to have you back on Languages Of The World in the future. If you would like to receive updates of our newest posts, feel free to do so using any of your favorite methods below: