Why the late Geoffrey Lewis’s Turkish Grammar (2d ed., Oxford, 2000) is exceptional:
At the beginning of a clause demek, demek ki, or demek oluyor ki (‘it becomes to say’) signifies ‘that is to say’: düşünüyorum, demek ki varım ‘I am thinking, which means I exist.’ (This Turkish translation of Descartes’ Cogito ergo sum is right—‘I am thinking’—and the usual English version—‘I think’—is wrong.)
I sent the foregoing to Facebook this morning, but this was not the best medium for the typographical features of boldface, italics, directional quotation marks, and indented quotations.
I had been aware that Lewis had died, but had not seen his Guardian obituary from 2008, which tells of a remarkable scholar born in 1920:
His written work divides sharply between the scholarly—on recondite topics such as Arabic studies of the philosopher Plotinus, Islamic surgery, Arab alchemy, and Turkish etymology and grammar—and books for more general readers written in easy-to-understand and often quietly droll terms. In 1953 he published Teach Yourself Turkish, for many years almost the only easily available work on the subject.
In 1955 came Turkey, a masterly one-volume introduction to the country, and in 1967 Turkish Grammar. Both books filled serious gaps in western European understanding. Although many later grammars have been written, Lewis’s is still regarded as the best.
I am sorry I missed the lectures that Lewis apparently gave in Ankara while I was living there. That I say “apparently” here shows the influence of Turkish, as learned from Lewis’s grammar. At the time of his lectures, I was not aware that they were happening. I know of them now only from hearsay—the Guardian article. In Turkish I would be grammatically required to indicate this. As Lewis explains:
32. Uses of the di-past. This is the tense used in speech when relating past events positively known to the speaker. If one has witnessed the arrival of a tourist ship, one may report the event in the words bir turist vapuru geldi. The newspapers will say bir turist vapuru gelmiştir, although in the headline they will use the synonymous but shorter geldi. Someone who has learned of the event from an eyewitness or from the newspapers will report it as bir turist vapuru gelmiş.
Thus for “Lewis apparently gave” I would say Lewis vermiş, because of what a newspaper would have reported as Lewis vermiştir. Lewis himself explains the subtle distinction as follows, where his first four examples translate as “came, saw, conquered, found”:
27. miş-past. This base is formed by adding -miş to the stem: gelmiş, görmüş, almış, bulmuş. Two distinct functions are combined in it.
It is first a past participle, describing present state arising out of past action. If you say kar yağmış ‘snow has fallen’, it may be that you yourself saw the snow falling, but that is not what you are concerned with. What you are reporting is not what happened but what is now the case: that there is fallen snow.
But precisely because the perfect participle does not indicate that the speaker has seen the action take place, it has come to be used as a finite verb to convey that the information it gives is not based on having witnessed the action but on hearsay or on inference from observed facts…
There is no inferential connotation when it is conjugated with -dir or with the past or conditional forms of the verb ‘to be’…
If you want to know about The Turkish ‘to be,’ see the Wikipedia article Turkish copula, which I created and which remains mostly as I left it in 2006. A knowledgeable person has done some work on the article since then, but somebody has marked it as needing work for improved clarity. The article still quotes the first part of a proverb whose totality is, Yolcudur Abbas, bağlasan durmaz “Abbas is a traveller: tie him down, he does not stop.”