Implicature

For the first time that I can remember, I have read a paper by a philosopher (reference 1) – a paper which was far more accessible than one expects such a paper to be. Prompted so to do by a chance encounter with a book (reference 2) by Martin Haspelmath, a German linguist who works on something called linguistic typology – having been sent to him by an email from Academia – people who are rather like Google in that once you show a bit of interest in something, they don’t let up. More about him in due course, but in the meantime I report on Grice.

The paper, published in 1975, opens by telling us about a dispute between people who believed that natural language, could be reduced, more or less, to mathematical logic and people who did not. A reduction which would, in theory, make natural language accessible to the same sort of algorithms and procedures which had been so successful in mathematical logic. There were also plenty of people who believed that words such as ‘and’, ‘or’ and ‘for all’ did not mean the same thing in the formal world of logicians as they did in the natural world of regular people. Grice argues the contrary case, introducing the device of the ‘implicature’ so to do.

A couple of his examples serve to give the idea. 

First, A is standing by an obviously immobilized car and is approached by B. The following exchange takes place:

A: I am out of petrol.

B: There is a garage round the corner.

Here, B is implicating A that the garage is open for petrol, or, at the very least, B has no reason to think that it is not so open. 

Second, A and B are just chatting, perhaps around the office coffee machine:

A: Smith doesn't seem to have a girlfriend these days.

B: He has been paying a lot of visits to New York lately.

Here, B is implicating that Smith has, or may have, a girlfriend in New York.

To support this process of implication, Grice introduces the cooperative principle, a principle which rests on the claim that one of the aims of conversation is to share information between the participants and that this works better if conversation proceeds along orderly, more or less predictable lines. The principle is organised into four categories – quantity, quality, relation and manner – and a number of subordinate maxims. A maxim under quantity is ‘do not make your contribution more informative than is required’. A maxim under quality is ‘do not say that for which you lack adequate evidence’.

One of the ideas here is that if one of the interlocutors breaks one of these maxims, there is going to be a reason and one of the tasks of the other(s) is to work out what that reason is. A task which is usually accomplished quickly and unconsciously.

Then we have it that a conversational implicature is something that can be deduced, given what is said and all the circumstances. A conventional implicature, possibly derived over time from a conversational implicature, does not follow so readily from what is said and all the circumstances and needs to be learned. Rather as idioms have to be learned by rote when learning a foreign language.

Not being explicit about all this stuff hovering in the background of the conversation has at least two advantages. First, the conversation does not get long winded, bogged down with all the detail. Second, inappropriately definite statements about matters which are uncertain or delicate are avoided. Tact often helps rather than hurts.

Grice goes on to build a taxonomy, a classification of implicatures, with their deduction, their extraction generally depending on knowing how people work and how conversations work, rather than on knowledge about the subject matter. Sometimes there might be more than one such deduction, possibly mutually exclusive, possibly not. There might well be ambiguity, possibly intentional. From all of which I associated to Empson’s famous book at reference 4, having a second life at about the time that Grice was writing, having been reissued in the learned Peregrine series from Penguin. Looking him up, I get to a list of Peregrines, including the different famous book at reference 5. Written, I believe, by the undergraduate tutor of the chap who taught English at my secondary school. In any case, the latter was very keen on the former and the book is now about to be mine, including packing and postage, for a fiver or so. Maybe I will make something of it.

What I have yet to think through is how this plays in text which is not dialogue. Perhaps in an essay about the part William the Conqueror played in draining the Lincolnshire fens, or about the role of the probation service in rehabilitating criminals. For example, one defect of offline text (as it were) is that there is no opportunity for amplification or correction when the reader is puzzled or on the wrong track, the sort of thing that can be picked up as one goes along online. Offline, one can use footnotes and parentheses, but these are a bit clumsy by comparison.

Then one can think of other aims of conversation. Perhaps to distract, perhaps to impress, perhaps as courtship, perhaps as contest; a war of words. Perhaps just as an elaborate game, perhaps the sort of game which was said to have been played at the French court in the run up to the Revolution. No noble aim, just a mutual desire for entertainment. Entertainment by means of clever and hopefully amusing conversation which does not depend on knowledge of beetles, operas or anything else. Entertaining froth spun out of nothing.

But none of this detracts from what Grice is suggesting. From the work that a computer is going to have to do if it is to compete with humans in the matter of comprehension – let alone conversation.

PS 1: both Grice and Haspelmath like to put important words into capitals, which I find visually irritating.

PS 2: Grice omits the now mandatory list of references. Perhaps being a book chapter derived or perhaps transcribed from a memorial lectures provides sufficient alibi.

PS 3: Grice mentions Shakespeare’s sonnet No.20 as including plenty of ambiguity, not so far removed from the sort of thing he is talking about. Reproduced in the snap above.

References 

Reference 1: Logic and Conversation – Grice, Paul – 1975. Reprinted from Syntax and semantics 3: Speech arts, Cole et al. ‘Logic and conversation’, pp. 41-58, (1975), with permission from Elsevier.

Reference 2: Indefinite pronouns – Martin Haspelmath – 1997.

Reference 3: https://www.academia.edu/. A for-profit outfit, name notwithstanding. But for-free they do you tell you about stuff which you would otherwise have probably missed. And make it available for download.

Reference 4a: Seven Types of Ambiguity – William Empson – 1930. Once owned in the Penguin Peregrine format but long since retired.

Reference 4b: https://en.wikipedia.org/wiki/Seven_Types_of_Ambiguity. 

Reference 5a: The common pursuit – F. R. Leavis – 1952. I associate to the time, perhaps fifty years ago, when I could have bought a great heap of ‘Scrutiny’ from a discount bookshop at the top of Neal Street for next to nothing. Must have been remaindered by the publisher. Possibly now an occult shop, although this last looks in Street View rather like a former public house. A further complication being that I remember this shop as being called ‘Poons’, now only known to Bing as the name of a small chain of Chinese restaurants, once known to us in Lisle Street for wind dried sausage and such like.

Reference 5b: https://en.wikipedia.org/wiki/The_Common_Pursuit. The wrong one.

Reference 5c: https://en.wikipedia.org/wiki/F._R._Leavis. The right one. I had forgotten that I went to the same secondary school as he did – although his father would have had to pay, which mine did not, the 11-plus having been invented in time. A chap whom I have noticed from time to time in the past, although not in the present fourth volume.