RT list: Submission from [Bertrand Labasse <B.Labasse-ScPresse@univ-lyon1.fr>]

Dan Sperber wrote :
>> My question is, why is this a problem, or if you prefer, what
>> problem is it?

Yes, the vagueness of effort and effect formulation can be a major
problem in many cases, for instance:
- to extend the relevance framework from an individual to a social
scale (example: social relevance of political news compared to
entertainment);
- to consider relevance in an applied perspective (example: didactics
of writing);
- to examine the compatibility (interface) between relevance theory
and the body of psychological knowledge on language processing
(example: text comprehension);
- to set up experiments with properly controlled variables, etc.

But it is not certain that jumping to neurology is absolutely
necessary. In fact, that would be quite an impressive jump, from an
heuristic conception to a positivistic conception of human cognition.
A middle way is to consider the tasks required in language
processing. I happened to work on a simplified taxonomy of these
efforts. Here is a simplified version of this already simplified
taxonomy (!), ranked from low level (bottom up) to high level (top
down) parameters :

- perceptive efforts (distinguishing signal from noise);
- lexico-syntaxic efforts (lexical access, sentence parsing);
- coherence handling efforts (inferences, etc.)
- mental modeling efforts (representation)

The point is that the two first levels of effort can indeed be
quantified. The third might be quantified, at the price of some
research. And I personally doubt that the fourth level could be
quantified, but I wouldn't bet too much on that. Anyway, four defined
kinds of factors (two of whom being quantifiable), is better than a
general concept.

When it comes to cognitive effects, the available literature is less
detailed, but we can assume that there are also four main kinds of
variables (Schank, 1979), ranking from low-level to high-level
factors. The lowest level includes stimuli common to most evolved
species - like sex, etc. By the way, I never figured out how this
kind of effect factors, apparently predominant in human information
processing (especially on the Internet...), can fit in the classic
definition of effect as "improvements to the individual's knowledge".
Anyway, here again, I am not sure that any of these classes of
factors could be accurately quantified, but they can be assessed.

Best wishes

    Bertrand Labasse

> At 20:36 13/10/2006, Stefan Malmberg wrote:
>
>> ...
>> In a recent article (Implicature vs.Explicature: What's the
>> difference?) (2006) Kent Bach writes :' The most obvious problem is
>> of how to quantify and measure degrees of cognitive effects and
>> processing effort. The formulations I've seen of
>> relevance-theoretic concepts and principles are too vague to be of
>> much help in this regard'.
>>
>> I too still find this a problem. Does anyone have any fresh
>> insights into the matter?
>>
>
>
> My question is, why is this a problem, or if you prefer, what
> problem is it?
>
> Since Relevance (1986), we have argued that relevance, effort and
> effect pay a role in comparing actually competing cognitive inputs
> or interpretations. The possibility of such local comparisons does
> not depend on an ability to quantify or measure these variable in
> the abstract. Then and since, numerous examples have been given
> comparing inputs and interpretations in terms of effort and effect,
> and experiments have shown how these could be used to derive and
> test predictions. In a more recent paper ("Modularity and relevance"
> in The Innate Mind: Structure and Content. Edited by Peter
> Carruthers, Stephen Laurence, & Stephen Stich, available at
> http://www.dan.sperber.com/relevance%20and%20modularity.htm) I have
> suggested how the mind cound indeed be geared towards the
> maximisation of relevance (i.e. the "Cognitive principle of
> relevance") on the sole basis of these local comparisons. In fact
> even "comparison" might be more than is needed, if by this we were
> to mean a cognitive operation of ranking alternatives. A
> non-cognitive sensitivity to gradients of effort and effect might be
> all that is required.
>
> So, you might say, maybe our minds, in their ordinary workings, just
> locally compare and don't quantify effect and effort, but we
> scientists still would like to be able to do so. If you are moved by
> this crave for measurement, then let me point out that what you
> should be after are neurological variables. Effort is just
> consumption of energy and should be measured in joules (how in
> practice, this is another matter). The measure of effect is more
> tricky and may have to do with spread of activation, which itself
> may have neurochemical correlates. What I don't see is why the fact
> that we don't know how to measure such variables should worry
> pragmatists, and particularly relevance theorists. If "how to
> quantify and measure degrees of cognitive effects and processing
> effort" is, as Bach claims, "the most obvious problem" with
> relevance theory, then let's cheerfully keep working on less obvious
> and more challenging problems in the field.
>
> Cheers, Dan
>
> -----------------------------
> Dan Sperber
> Institut Jean Nicod
> http://www.institutnicod.org
> 1bis avenue de Lowendal
> 75007 Paris, France
>
> email: dan@sperber.com
> web site: http://www.dan.sperber.com
> -----------------------------
Received on Sun Oct 15 22:41:18 2006