I spoke to a clinical psychologist recently about how verbal subtests on IQ tests are evaluated. She said that answers regarding the meaning of a word were evaluated based on whether or not they were the same as what most people answered. I assume an explanation that is nearly identical to the normal explanation, except for different but synonymous words being used, is counted as correct. Now, this struck me as very weird. I assume the logic is the following:
Words are tools of communication. Using words as if they mean something most people do not think they mean is a deficient use of the tools; assuming people actually want to communicate as best as possible, if they use words in an atypical way, it implies they have a low comprehension of how other people use those words. This is thus indicative of their reasoning abilities.
I find the above reasoning mildly problematic in itself, but it is its application in this context that I do not understand at all. While I agree that atypical usage of words often implies poor comprehension of other people's usage of words, I do not agree that atypical explanations/definitions of words implies atypical usage of words. Why? Because common sense fails all the time. Let's say the test asks you to define table. What is the first thing that comes to mind? To me, it is:
A surface with four legs.
Very sensible, right? As a definition though, it is quite deficient. It does not proficiently represent my, or other people's, usage of the word. Why? Because a lot of objects most would identify as tables are left out, and a good amount of things that most would not identify as tables are defined as tables by that definition. Here's a better definition, that might be a bit atypical:
A balanced surface with some number of legs, and with a height at the level of a seated person's torso.
The second definition might be a bit less natural, but it definitely represents usage far better. Now, three-legged, six-legged, single-legged, etc., tables are included, and structures resembling that of a table, but that are too tall for use, are excluded. For some people, function is also quite important in definition, and as such, one could include a criterium of load capacity in some range as well. I don't know if my given examples are very good. Perhaps the first definition is far too naïve for most people? Perhaps the second definition isn't that unnatural? I don't know, but I think you get the general point I am making. The more intelligent you are, the more you can grasp the complexities of our messy words; thusly, the more intelligent you are, the more atypical your definitions would likely be (if the goal is to represent typical usage that is).
So, how then can the verbal subtests be evaluated by checking how well the answers correspond to typical responses? That sounds absurd to me. If the question is; "define this word in the way you reckon most people would", then I get it (though in that case I'd say the items are testing social intelligence more than verbal). But if the goal is to define a word, then the goal is to represent its usage in a way that includes the most examples and excludes the most non-examples. Which brings me to my question: how are the verbal subtests on IQ tests evaluated? I am both interested in the questions asking for the meanings and relations between words, as they both come down to the same thing, ultimately. Do you know, or have any helpful references on the topic?
