If I Had You: The Implicature

Grice _knew_ that 'have' as per 'capio' was a rude auxiliary.

"I have eaten my sandwich"

In Latin, "I have my sandwich qua eaten".

What especially irritates a Christian Gricean is the ambivalence of "The lord is risen today" -- surely that's correct, and no need, as we hear sometimes, "the lord HAS risen today".

"Has" is not only rude but a Casanova of a word.

But we do care for Carey.

In her

"The role of conversational implicature in the early
grammaticalisation of the English perfect"

-- Being Essay III in Hall K. (Ed.)
_The Legacy of Grice_. Berkeley: Berkeley Linguistics Society.


she proposes a serious Gricean study of the "ambiguity", as we may call it, of the English perfect.

Consider:

1. "I have understood the message", said the King.

Does this mean that _he_ understood it? Carey thinks this is cancellable.
e.g. as in

2. The king has the message understood by the scribe.

In other words, the perfect was originally periphrastic.

Carey's interpretation is is Gricean in that it's guided by considerations
of conversational relevance, as in his famous cooperative principle of
conversation:

Make your contribution such as is required at the stage at which it
occurs by the accepted purpose or direction of the talk exchange in which
you are engaged.

CATEGORY OF QUANTITY.

1. Be as informative as is required.

2. Do not be more informative than is required

CATEGORY OF QUALITY:

Try to make your contribution one that is true.

That is,

1. Do not say that which you believe to be false.

2. Do not say that for which you lack adequate evidence

CATEGORY OF RELATION.

Be relevant

CATEGORY OF MANNER

Be perspicuous.

That is,

1. Avoid obscurity of expression

2. Avoid ambiguity

3. Be brief (Avoid unnecessary prolixity)

4. Be orderly.

which with WoW:xvii gives us ten commandments (Grice's decalogue).

In the 'Introduction', Carey refers to work by E. Traugott and L. Horn.
Semantic change is brought about by conversational implicature through
"strengthening" (informativeness constraint).

Traugott has shown how this explains the change from the temporal to the
concessive meaning of "while" in English, and also the shift from the
deonitc to the epistemic meaning of "must".

Carey also quotes Horn on lexical narrowing and broadening. She writes: "The early grammaticalisation of the perfect is yet another example of the role of conversational implicature in instigating semantic change."

In 'Prolegomena' Carey considers the shift from temporal to causal reading
of "siththan". When
"siththan" was still used as a temporal connector, there are instances of
"causal" implicatures:

e.g.

(1) Tha siththan he irre waes & gewunded, he ofslog
micel thaes folces. (Orosius).

Then, since/after he was angry & wounded, he
slaughtered much of that troop.

Carey writes:

"Traugott notes that "irre" favours the _causal_ reading if
"the addressee assumes that the utterer is obeying the Gricean maxim of
Relation ('be relevant')."

Today, a purely causal reading is possible:

(2) Since you are not coming with me, I will have to go alone.

What does Onions say about this. His entry for 'since' goes:

since: ME synnes,

reduced form of "siththan" i.e. "sith": after, "than".
cfr. G. seitdem, since.
rel. to Latin, "setius".

In section 2 ('Synchronic facts about the English perfect'), Carey refers
to the "seemingly schizophrenic nature of the English perfect", and draws
on a taxonomy of 'usages' of the perfect from Brinton:

Type i. RESULTANT STATE

Past Action has persent result:



He has caught a cold (+> can't work)
I've eaten lunch, thanks (+> I'm not hungry).



(were "+>" indicates "conversationally implies")

Type ii. Past situation continues up to the present.

I have read that novel
I have been abroad several times.

Type iii. A situation began in the past and persiss today:

We have known him since he was a child.
He has sung in the choir for years.

Type iv. The Perfect of Recent Past.



There is a kernel (core, yolky) meaning to all this divergence of meaning, Carey thinks.

To draw from L. Bauer: "an accomplished fact which is not conceived as
separate from the moment of speaking".

Carey further notes Bauer's classification of verbs as 'telic' and
'non-telic' which she finds to be relevant as to the conventionalisation
process.

In the section 3 ('The diachronic development of the English perfect'),
Carey quotes from Mitchell, who notes:

"in the original form, the perfect periphrasis was adjectival and
inflected".

The modern equivalent would be a 'stative (vs. achievement) construction'
as in

I have the car washed.

rather than

I have washed the car.

While in _modern_ English the order (of object and participle) is crucial,
this was not so in OE.

(2) Fela Godes wundra we habbath gehyred & eac geswene (AECHom)
We have heard and also seen many of God's wonders.
(geswene is inflected; gehyred is not).

Carey notes one textual problem here: "the earliest stages of the
development of the perfect predate the OE texts".

Carey considers "contextual constraints on the first perfects".

"Predictably, the first uses with the perfect meaning emerged from contexts
in which the adjectival construction was still operative." Carey considers

*I have my room cleaned but you messed it up.
(Because "I have my room cleaned" +> it is _still_ cleaned).

(were "*" indicates "ungrammaticality").

She considers a perfect referring to multiple resultant states:

Grandma has baked a pie every Saturday for the last 20 years.

But notes,

"It's highly unlikely that the first uses of the perfect would [in the
History of English] refer to multiple rather than single resultant states."
She also considers:

* I have the Italian shoes wanted.

(an atelic verb is incompatible with the adjectival construction).

Of the 167 OE examples from King Alfred (84 instances) and Aelfric (83
instances) which Carey draws from Venezky and DiPaolo) Carey finds only _2_
instances with atelic verbs. She hypothesises that the first use is then as
a resultant state: a past action with a present result.

Carey then considers the shift from the adjectival to the first perfect
construal. Are we talking of different "truth-conditions" here? Carey
thinks not:

"Importantly, these differences between the adjectival construction and the
first resultant state perfect do _not_ correspond to different truth
conditions but to different _construals_ of the _same_ truth-condition."

There _is_ a truth-conditional difference, though: in the adjectival
construction the subject may _not_ be the _agent_. In contrast, in the
perfect construction, the subject _must_ be the agent.

(3)

A: John has bathed Meow. Meow's ready for the cat show now.

B: John bathed Meow?

A: *No. Mary did.



This is a matter of truth conditions, and not implicatur. For if it were an
implicature, it would be cancellable, and it's not.



Carey also considers:

(4) Ic haebbe gebunden thone feond the hi drehte (Aelfric, Hom i 31).

This has two readings. The first is truth-conditional:

i. I have that enemy bound that afflicted them.

The second is implicatural:

ii. I have bound that enemy that afflicted them.

The idea is that implicatures which are merely 'contextual' become
'indexed' and then incorporated into the 'meaning' of a given form. But how
is this done? This she considers in the next section.

In 'The role of mental state verbs in the conventionalisation process',
Carey considers the sentence:

"Now that I have the riddle understood,
I can start on the next one, I guess."

What is relevant there is the change of state of the subject, not the the
understood state of the riddle.

With "that-clauses" following the verb the shift from the adjectival to the
perfect construal is even more apparent":

(5) Ic haebbe nu ongiten thaet thu eart gearo to ongitanne mina lara.
(King Alfred. Bo. 36)
I have now understood that you are ready to understand my teachings.

There are also uses with verba dicenda:

(6) Nu haebbe we ymb Affrica Landgemaero gesaed. (Or. 1.)
Now we have talked about the african land.

"Verbs of reporting are similar to mental state verbs by typically
including a conversational implicature in which the participle has a verbal
rather than an adjectival function."

In 3.4 Carey considers further "Evidence from Old English".

Data from the 167 instances suggest that once the mental state and the
reporting verbs have incorporated the perfect meaning as a legitimate
meaning of the "have + participle" construction, it can start to extend to
other verbs. 32% were mental state verbs, 19% wre reporting verbs. 40% were
verbs which designated a change of state in the object, and the remaining
9% verbs which do not designate a change qua perceptually verifiable.

"It is clear that by ME, the perfect had extended not only to atelic but to
stative verbs as

(1) His hert has ever ben the with (Curs M 7994)

Also, by ME examples of the perfect of experience had emerged:

(2) Your owne knight, that serveth you with wille, herte and might,
and ever hath doon. Chaucer CT A3079).

Carey concludes her essay with the observation that the first uses of the
perfect were resultant state perfects with telic verb participles. A time
adverbial may be crucial. "All of the early examples with activity verbs
included a time adverbial. The addition of the time advebial provides
an implied goal, so that the event is telic although the verb is not."

There seems to be a general principle underlying here:

"A conversational implicature associated with a grammatical construction is
more likely to be conventionalised if a subset of the uses of that
construction truth-conditionally or logically or semantically (rather than
pragmatically) _entails_ some part of the new meaning".

"Recall that in the case of the perfect, the uses with mental state verbs
entailed that the subject was the agent of the process."

So, what HAVE _you_ done to deserve this?

---

Etc.


References

BRINTON L. The English Aspectual Systems. CUP.
CAREY K. The role of conversational implicature in the early
grammaticalisation of the English perfect. In K. Hall, The Legacy of Grice.
"Respectfully dedicated to the memory of H. P. Grice". 16 essays in memory
of H. P. Grice contributed at the 16th annual meeting of the Berkeley
Linguistics Society.
COLE P. Synchronic and diachronic aspects of conversational implicature. In
Cole.
GRICE HP Studies in the way of words. Harvard University Press.
GRICE HP Logic and conversation. Repr. in A Kasher, Implicature.
HORN LR A taxonomy of pragmatic inference: q-based and r-based implicature.
In Kasher.
ONIONS C. Oxford Dictionary of English etymology.
TRAUGOTT E. On the rise of epistemic meaning: subjectification in semantic
change. Lg 65.
TRAUGOTT E OE syntax in R. Hogg, The Cambridge history of English.
TRAUGOTT E. The pragmatics of grammaticalisation. In Traugott, Approaches
to grammaticalisation. Benjamins.