Ambiguity

Barner and Stiller both discuss the notion of scalar implicature in relation to the learning of language and its processing in children versus in adults. Whereas Barner focuses more on the nature of scalar implicature and the complementary existence of scalar alternatives in learning in children, Stiller extends the discussion to the importance of property base-rates in how adults probabilistically integrate world knowledge with linguistic and social factors.

Gricean pragmatics stipulate that a helpful and cooperative speaker would make statements that are as informative as needed, and that convey no more information than is presently required. From these assumptions stems the conclusion that any more informative statement than that which was given must contain some falsity, as some information in the new statement would not belong in the subset of the original. Therefore, equivalence classes exist in a structure of scalar implicature strength – for example, ‘and’ statements are stronger than ‘or’ statements, while in propositional logic, Truth and Falsum are the strongest and weakest, respectively.  

This leads us to consider a closely-related topic: the universal and existential quantifiers in first-order logic. As discussed in Philosophy 150, ‘some’ has a different meaning when pragmatically defined as it does when logically defined. When does ‘some’ entail ‘all’? Likewise, adding two negations before a propositional atom ought to be equal to the propositional atom itself. For example, ‘the movie is not not interesting’ should logically mean that the movie is in fact interesting. However, with linguistic inflection and emphasis, the sentence could be articulated as ‘the movie is not not interesting…’, hinting that it is not entirely interesting, but is not entirely boring either. Thus, natural language and the language of logic aim to represent the same ideas, although the latter does so with more precision but far less nuance.

I feel that such a difference complicates the notion of scalar implicature. The readings imply that there are at least two different definitions of the word ‘some’, for example. One is logically tight, the other pragmatically so. If we were to additionally consider alternatives not only within one scalar set (‘some’ vs. ‘all’) but also entire alternative scalar sets (<one, two, three> versus <a, some, many, most, all>), nuances and thus ambiguities are further exaggerated. I feel that this would further complicate learning in children: as mature adults, we may believe that we have a clear notion in our heads of what ‘some’ means. However, not only is the notion of ‘some’ not even fully innate to begin with (since children need to learn the concept) – it seems to have an ever-morphing definition depending on situation and context.

Language aims to make it possible to express oneself in as precise a set of terms as possible, in order to communicate ideas to others. However, it remains highly ambiguous at its core, even with the introduction of notions such as ‘scalar implicature’ which attempt to define languages in a rigid and structured manner.