(Fwd) even if

[And Rosta <a.rosta@UCLAN.AC.UK>]

Some people were wondering about "even". Here is a message
about it from another list.

--And.

------- Forwarded Message Follows -------
Date:          Mon, 20 Oct 1997 15:04:11 +0200 (METDST)
From:          Delgado Lavin Eva <fipdelae@vc.ehu.es>
To:            relevance@phonetics.ucl.ac.uk
Subject:       even if


Dear relevance listers:
I have been thinking about 'even' (among other focus expressions) for
years as part of my PhD research; in fact, I am really interested in
accounting for 'even-if' conditionals, but this can only be done if one
understands what 'even' and 'if' contribute separately to the compound
'even if'.  My contribution to the list today will focus on 'even'.
        As J. Bennett pointed out in his now classical (1982) paper on
'even', an even-utterance like the one below communicates three sorts of
assumptions (a rough paraphrase of each of which is given):

Even Max tried on the pants
A:  Max tried on the pants
B:  At least one person other than Max tried on the pants
C:  The fact that Max tried on the pants is, or is purported to be, more
surprising, less likely, expected or whatever, than this other
person(persons)'s trying on the pants.

The question is: what is the contribution of 'even'?  My idea is that,
like any other focus expression (understood widely to cover contrastive
stress, focus adverbs and certain syntactic constructions), 'even' evokes
what I have chosen to call an 'assumption schema' (it is in fact a
grammatically specified entailment, but the term 'assumption schema' is
more graphic, I think; if I am using the term wrongly, please somebody
correct!) with a slot or variable in the place of the focused
constituent.  This schema (or entailment, as you wish) can be relevant in
two ways: to take the example above, if the hearer has a number of people
in mind, the schema can be filled (or the variable instantiated) -e.g.
you could obtain other similar propositions 'Mary tried on the pants',
'John tried on the pants'-; if the hearer does not have anybody in
particular in mind, as is the case with us reading the utterance out of
context, for the evoked schema or entailment to have some relevance it
would be assumed that the variable in it has some value different from
Max.  So far, we have:

Utterance:  Even Max tried on the pants
Explicature:  Max tried on the pants
Schema or entailment:  Someone tried on the pants
+ If one or more instantiations prove relevant, several propositions
(e.g. 'Mary tried on the pants', 'John tried on the pants'...)

So B above is communicated by virtue of 'even' being a focus particle (an
utterance with 'only' or 'too' or contrastive stress would communicate
the same in the same way).  But each focus expression has its own
particular contribution that distinguishes it from other focus
expressions. I claim that 'even' helps in the selection of the context.
It 'instructs' the hearer that the context for processing should contain
an assumption with the following property: it is the contradictory of the
explicature. I claim that this assumption is manifest in the context in
the sense that it is derivable (from encyclopaedic assumptions which are
then also thereby evoked). In our example, 'even' would tell the hearer
-us- that the assumption 'Max did not try on the pants' is manifest (and
so derivable from certain encyclopaedic assumptions about Max and how he
behaves in certain situations, etc., whatever this is, we can't tell
because we do not have any information about Max; however, it is
interesting that even without knowing or caring about this Max, we are
forced to assume that he is shy, or does not like that kind of situation,
or something like that, so we form certain assumptions regarding him in
interpreting the utterance).
        It is important that 'even' instructs us to use this
contradictory manifest assumption as part of the context, because thanks
to this, the manifest assumption is sure to be eliminated (since the
explicature of the utterance, coming from a trustful speaker is
presumably stronger than an assumption derived from encyclopaedic
assumptions). Notice that if 'even' is absent ('Max tried on the pants')
the speaker cannot ensure that the hearer will eliminate the manifest
assumption ('Max did not try on the pants'), because perhaps it will not
be derived from encyclopaedic information to begin with.
        It goes without saying that 'even' is not required for the
elimination of erroneous manifest assumptions. An 'even' utterance is
just a more explicit version for what the speaker means.  Take the example
from Lycan (1991) (1) and my analysis (2):

(1) It's fast, it's durable, it's even low in price
(2) a. 'Even' utterance: It's even low in price
    b. Explicature: Product 101 is low in price
    c. Schema or entailment: Product 101 is property x
    d. Manifest assumption: Product 101 is expensive

'it's expensive' is manifest because it can be inferred from our
encyclopaedic assumption 'Quality products are expensive' and the new
assumption 'Product 101 is a quality product' (derived in turn from the
previously communicated assumptions 'product 101 is fast' and 'product
101 is durable'). Now the explicature 'product 101 is low in price'
contradicts the manifest assumption and, being stronger, eliminates it
(causing in turn probably the revision of the hearer's encyclopaedic
assumption 'quality products are expensive' or perhaps of the assumption
'product 101 is a quality product'). All this processing 'could' happen
without the presence of 'even' but it is obligatory when 'even' is used.
        For 'if' I follow S. Barker (1994).  According to him, in 'if P,
Q' 'if' signals that P is being supposed (i.e. entertained as true). In
asserting a conditional, the speaker is committing himself to the
truth of Q and is signalling that this commitment holds under the
supposition of P. So to take Simon's example

a. Even if I study hard I will fail
b. Explicature: SUPPOSITION (speaker studies hard), the speaker will fail
c. Schema or entailment: SUPPOSITION (...), the speaker will fail
d. Manifest assumption: SUPPSOTION (speaker studies hard), the speaker
                         will not fail.

Notice that the manifest assumption, which the speaker wants to ensure
will be eliminated by the explicature of his utterance, is derivable from
the encyclopeadic assumption 'if you study hard for an exam, you will
pass it' (or something along those lines). Notice, too, that the slot in
the schema (or the variable to be instantiated) in c. will have a highly
accessible relevant instantiation: the negative version of the antecedent
used in the utterance. In this way, a conditional is obtained that says
that if the speaker does not study hard, the speaker will fail. Putting
this together with the explicature of the utterance, the effect is that
the consequent follows.  However, I do not think it is fair to say that
it is entailed (and there are examples that show that 'even-if'
utterances do not entail their consequents, although very often they seem
to).
        I hope this was clear enough for Simon and the rest of you, and
of course it is all my responsibility if I have used relevance-theoretic
notions wrongly. Please do let me know what you think.

Best wishes,
Eva

Eva Delgado Lavin
Departamento de Filologia Inglesa
Facultad de Filologia
Universidad del Pais Vasco
Paseo de la Universidad s/n
Apartado 2111
01006 Vitoria-Gasteiz (Spain)

<fipdelae@vc.ehu.es>