[hist-analytic] Shuga-Free

Roger Bishop Jones rbj at rbjones.com
Fri Feb 12 15:31:57 EST 2010

On Tuesday 09 Feb 2010 20:55, Jlsperanza at aol.com wrote:

> "If we are  looking for a possible FORMAL COUNTERPART
> of such a sentence  as
> [the author of 'Sein und Zeit' wears a  moustache]
> we have TWO candidates to consider."
> [the first  candidate being]
> (ix.Zx)Mx
> "in which the  iota-operator is treated as being
> -- the second candidate being
> M(ix.Zx)
> "in which the iota operator is treated as a
> DEVICE for  forming a TERM"
 > -- and not as a 'term' itself, as I wrote in my previous  post, which would
 > be stupid.

Well actually, not so stupid.
In HOL the choice function _is_ a term as are all the variable binders;
but this is unimportant.  Lets suppose we are working in a first order 

These are two alternative syntaxes, and there is more than one
plausible reading of the semantics for each syntax.
Did Grice say anything at this point about the semantics.

Lets proceed under the assumption that the second should be
read as the most direct interpretation of Russell's theory of

    ((ix.Zx)Mx) <=> (Ex)(Zx /\ (y)(Zy => y=x) /\ Mx)

> Grice continues with a very  important point -- which he then owes to
> 'Shuga':
> "If we select  [the first candidate], then, when we introduce  negation,
> we have two  SEMANTICALLY distinguishable ways of doing so"
> a.1.   - ((ix.Zx)Mx)
> a.2     (ix.Zx)-Mx
> "The second [a.2] will, and the  first [a.1] will not, ENTAIL
> the existence of an x [e.g. Heidegger] that is  UNIQUELY
> [the author of Sein und Zeit]"
> "But if we select [a.2] there  is only ONE place -- prefixing --

I think you mean here [the second candidate].

> for the introduction of  negation."
> "And, in consequence,
> (ix.Zx)-M
> -- ["The  author of Sein und Zeit did not ALWAYS wear a moustache"]

Surely not.
If the semantics were as I mentioned above then this should be the obvious 

> -- ["The  author of Sein und Zeit did not wear a moustache"]

> "will be a  [scope-] ambiguous structure"
> "unless we introduce a disambiguating scope  convention"

I don't see why.
Before we rendered Russellian description as a quantifier we really did
have a scope ambiguity, but once it is a quantifier the scope rules for
quantifiers sort it out for us.

> -- as Grice does: two of them: the square-bracket device,  which turns
> certain expressions immune to negation, and, in "Vacuous  Names", the
> numeral
> subscripts.
> Grice continues:
> "One  ONE REDUCTION to primitive notation

Is that "On ONE"?

> the existence of a unique [author of Sein  und Zeit]
> will be ENTAILED"
> -- the term is Moore's.
> [and thus  we do not need Strawson's notion of 'presupposition']

Which was always surely an alternative to Russell's theory rather than
an interpretation of it.

> "On the other will  will not", but will be 'implicated' conversationally;
> and thus we  don't need the Strawsonian notion of truth-value gappy
> presupposition  either -- but cfr. Noel Burton-Roberts and his
> neo-Strawsonian
> fanfares.

I agree that neither of the two alternatives "needs" Strawsons gaps,
they are both ways of doing without them and therefore staying within
a two-valued logic.

However, the point of Strawson's gaps is not that we need them,
but that they provide a better account of the way natural languages
work.  This is a matter of controversy, but it does not address this
issue to argue that there are alternative ways in which descriptions
can be understood.

> Grice continues:
> "Call these respectively the STRONG  and the WEAK readings."
> How to decide? Grice is clear: his ear for  English:
> "Now, if there  WERE a clear distinction in SENSE [Fregean  sense]
> (in English) between, say"
> S   [the author of  Sein und Zeit did not wear a moustache   then]
> and
> W   [it is not the case that the author of  Sein und Zeit  wore a moustache
> then]
> "(if the former demanded  the STRONG reading
> and the latter the WEAK one), then it would  be
> REASONABLE to correlate"
> 'the author of Sein  und Zeit wears a moustache"
> "with the formal structure that treats  the
> iota operator LIKE A QUANTIFIER"
> ('as a quantifier' seems better  English -- cfr. P A Stone on 'Girls won't
> be girls')
> "But this does NOT seem to be the case; I see
> no such clear  SEMANTIC distinction"
> nor do I. Geary sees, but a pragmatic distinction,  and then, not _that_
> clear -- "On a clear day you can see forever" --  "on a VERY clear day you
> can
> see as far as Aylesbury" Noel  Coward).
> Grice continues:
> "So it seems BETTER to  associate"
> the author of Sein und Zeit wears a  moustache
> "with the formal structure that treats the iota operator
> as  a term-forming device"
> (and not as a term simpliciter as I clumsily wrote  in my previous).
> "We are then committed to the STRUCTURAL  ambiguity
> of"
> the author of Sein und Zeit does not wear a  moustache
> [eg. since he died some years  ago]

How is this a structural ambiguity?

> Having credited Shuga with that Grice finishes the   consideration:
> "The proposed task may NOW be defined as follows: on  one  reading, [the
> author of Sein und Zeit does not wear a moustache]  entails the  existence
> of a
> unique [author of Sein und Zeit], on the  other it does not; but  in fact,
> without waiting for disambiguation,  people understand the utterance  of
> [the author  of Sein und Zeit does not wear a  moustache]
> as IMPLYING (in some  fashion) the unique existence of [an author of Sein
> und Zeit]. This is  intelligible if on one reading -- the strong one -- the
> unique  existence of [an author of Sein und Zeit] is ENTAILED, on the other
>the  weak one -- though NOT ENTAILED, it is *conversationally implicated*   
>[emphasis mine. JLS]"
>"What needs to be shown [he does], then, is a  route by which the weaker  
>reading coult come to implicate what it does  not entail"

However, it seems to me that the problem is with the truth conditions for
the sentence.

The merit of Strawson's treatment as presupposition is that we then get out
of having to decide on the truth value in the problematic cases.
But talk of implicature doesn't solve that problem.

If we have to assign a truth value when unique existence fails, then it
is natural to chose "FALSE" (and we then get the problem of what we are
assigning false to if we don't have the disambiguating quantifier syntax).
However, you cannot chose the truth value independently of your choice
about what is implied or entailed (these are the same in first order logic 
since it is complete) by the construct.

If you chose FALSE then the sentence will imply unique existence,
whether you like it or not (assuming a two-valued logic).
So then it becomes superfluous to invoke implicature.
Chosing TRUE is even worse, for then the description becomes pretty

So IF it was Grice's aim to argue the superiority of implicature over 
presupposition, the argument if I have it right, fails.


More information about the hist-analytic mailing list