Relevance_Archives: Re: RT list: Re: RT List: the conceptual-pro

From: Christoph Unger <christoph-kuelvi_unger@sil.org>
Date: Sat Jan 05 2008 - 14:32:10 GMT

Dear Minh Dang,

On Fri, Jan 04, 2008 at 05:00:42AM -0800, Minh Dang wrote:
> ...
>
> 2. Criteria of distinction
>
> For the distinction to be of real substance, it is necessary to establish
> some criteria to distinguish expressions that encode concepts from those
> that encode procedures. RTheorists have made several attempts in this
> direction (Wilson and Sperber 1993, Blakemore 2002, 2006, Iten 2004). In
> what follows, I will summarise those criteria proposed and we will see if
> these criteria are sound or reliable.
>

Notice that these criteria describe properties that linguistic
expressions encoding procedural meaning may be expected to have if the
theoretical claims about procedural encoding are true. There is no
claim that all expressions encoding procedural meaning should exhibit
all these properties. Nor is there any claim that some or all of these
criterias are failsafe identifiers.

This is because the real question is not so much about criteria
itself, but about the theoretical content of the notion `procedural
encoding'. Unfortunately, we still seem to have a better idea about
what procedural meaning is _not_ (e.g. not conceptual encoding, not
translational encoding) than about what it is, other than
metaphorical proposals such as `taking the hearer part way' to the
intended interpretation. To me this is more worrying than the fact the
proposed criteria are not failsafe (and not applicable to every
linguistic expression).

Best,
Christoph

-- 
Dr. Christoph Unger
SIL International 
Alleestr. 7
67308 Albisheim
Germany
Phone: +49 6355 989939

Received on Mon Jan 7 13:54:55 2008

This archive was generated by hypermail 2.1.8 : Mon Jan 07 2008 - 14:07:53 GMT

Notice that `so', `but' etc. do not figure in the mental
representations communicated (be they explicatures or
implicatures). So how could they have logical properties in the mental
representations communicated?

>
> In general, it seems the claim that only conceptual expressions have
> logical properties whereas procedural expressions cannot be maintained. A
> procedure, an instruction, a guide must have its contents. It would be
> very odd to assume that this content have no logical properties yet still
> fulfilling its guiding or constraining function. To claim that procedures
> do not have logical properties is in effect to allow that a procedure can
> be something like ‘do A and don’t do A’. It is not clear how such a
> non-logical procedure/constraint can do the guiding/constraining job it is
> supposed to do. In other words, a procedure which have no logical
> properties would give contradictory instructions which confuse rather than
> guide the hearer’s interpretation. In addition, if we accept that a
> procedure/instruction must have some sort of content, this content must be
> conceptual by nature.

> It would be a fantastic miracle if a conceptually
> empty linguistic form could do the guiding or constraining
> job.

Not at all. The linguistic expression could be linked to a cognitive heuristic
(or even an algorithm). For example, `so' may activate a procedure
that results in increasing the activation level of all mental
representations that have the explicature of the clause introduced by
`so' as the second part in a 'if P then Q' schema, where this
procedure starts in working memory and may spread through all memory
regions having connections to representations in working memory until
processing effort gets too high. (This is pretty much the standard
account of `so' in Blakemore 1988). This is certainly not a
concept. It is true that I have to use coneptual expressions in
English in order to explain how this procedure works; but this doesn't
mean that the procedure has conceptual content.

Of course, such a procedure or heuristic or algorithm does have
logical properties that make it possible to be computationally
implemented in some device. So in this sense one may be able to say
that `so' has the logical properties that describe the procedure. But
surely these logical properties are not properties of any communicated
mental representation, hence they are not conceptual information.

This archive was generated by hypermail 2.1.8 : Mon Jan 07 2008 - 14:07:53 GMT

Re: RT list: Re: RT List: the conceptual-procedural distinction