Our capacity to disambiguate

Perhaps the most infuriating riddle I've ever come across reads as follows: Two coins together sum to a value of 30 cents, and one of them is not a nickel. What are they?

The incredibly annoying answer is that they are a quarter and a nickel. One of them isn't a nickel, but the other quite emphatically is.

Granted, this riddle has never actually been posed to me. I came across it in an episode of the tv show Scrubs, in which someone is so incensed by the riddle's answer that he destroys the asker's bike. As amusing as this was to watch, it also got me thinking about ambiguity in communication, and how in general, we're able to understand exactly what someone means, even when their utterance is logically suspect. 

Now, 7 weeks into Linguistics 1 and after a year of studying Symbolic Systems, I can formalise my initial epiphany.  I know that our ability to accurately surmise what our conversation means is modelled by Grice's maxims and the cooperative principle. I know that the demographic background of who I'm speaking to plays a crucial role in my ability to discern what it is they are saying, and its veracity, as does my level of exposure to their dialect and diction. This weekend's readings gave a new dimension to this perspective, and one I feel filled the last major gap in my rudimentary understanding of how we use language to communicate information. 

Stiller's paper provided a fascinating insight into the mechanics of scalar implicature. I would have expected that the Gricean counterfactual paradigm would hold here, and that children would be able to make appropriate distinctions given purely the logical implications of an utterance. The paper quickly disillusioned me, however. Even I wasn't entirely certain that if I was a subject in experiments 2 and 3, I'd be able to ascertain specifically which object the sentence referred to. The discussion revolving around information theory - how "an informative expression literally conveys more bits of information about which object is being talked about within a context" seemed to me the most interesting part of Stiller's argument, partly because information theory is something I'd like to learn more about and partly because the relationship between the salience of a feature and our ability to discriminate based on it seemed so obvious, but wasn't something I'd have ever thought of. At the end of the paper, it seemed to me patently obvious that scalar implicature was contingent upon real world contextual knowledge.

I felt that Barner's paper  - while interesting - presented a considerably less compellingly unique argument. The inability of children to calculate scalar implicature or to conceive alternatives to scales like some or all did not seem an especially revelatory insight. After all, children might simply lack the experience to understand that the world doesn't know what they're thinking, or the vocabulary to specify precisely how much of a thing they are talking about. In fact, it seems to me that the paper telling us that when alternatives are specified contextually, children complete the task without difficult. I failed to see how those results could speak to our capacity to deal appropriately with scalar implicatures so much as they did to the ability of children, which is a far weaker conclusion.