Socrates' “strangeness” is the keynote of Alcibiades' speech about him in the Symposium. The talk starts on that note (215A); and reverts to it near the end:

This book is for readers of Plato's earlier dialogues who have felt this strangeness, have asked themselves what to make of it, have pondered answers to its enigmas, and are willing to work their way through yet another. What I offer should not distract them from their encounter with the Socrates who lives in Plato's text. It should take them back there for a closer look.

The book has been a long time in the making. It started with a non-start. A stroke of luck in 1953 had assured me of a year entirely free from teaching. I had gone to the Institute for Advanced Study at Princeton to work on Plato's philosophy, uncommitted as to how I would proceed. Harold Cherniss, the great scholar in my field there, declined to give advice – how like Socrates he was in this. Left to my own devices, I cut the project in half, allocating the whole of that year to a study of Plato's Socratic dialogues.

Socrates' commitment to reasoned argument as the final arbiter of claims to truth in the moral domain is evident throughout Plato's Socratic dialogues. He refers to it in the deliberation by which he justifies to Crito the decision to remain in prison and await execution:

And yet he is also committed to obeying commands reaching him through supernatural channels. When explaining at his trial why the state's power of life and death over him could not scare him into abandoning the public practice of his philosophy, he declares:

Between these two commitments – on one hand, to follow argument wherever it may lead; on the other, to obey divine commands conveyed to him through supernatural channels – he sees no conflict. He assumes they are in perfect harmony. Can sense be made of this? I want to argue that it can.

In the last and most famous of his Theses on Feuerbach Marx observes: “The philosophers have done no more than interpret the world. The point, however, is to change it.” Substitute “morality” for “world” and the observation would be true of almost all the leading philosophers of the West. Moralists as powerfully innovative as are Aristotle, Hume, and Kant take the morality into which they are born for granted. The task they set themselves is only to excogitate its rationale. It does not occur to them to subject its content to critical scrutiny, prepared to question norms ensconced in it which do not measure up to their rational standards. But there have been exceptions, unnoticed by Marx, and of these Socrates is the greatest. Proceeding entirely from within the morality of his own time and place, he nevertheless finds reason to stigmatize as unjust one of its most venerable, best established, rules of justice.

By the morality of a society I understand those norms of right and wrong, rules of conduct or excellences of character, publicly acknowledged within it, whose function it is to foster human well-being. The sense of justice centers in the concern that those norms be applied impartially.

To single out one of the many values in our life, elevating it so far above all the rest that we would choose it at any cost, is one of the many things that have been called “romanticism” in the modern era. Its typical expression there is sexual love. To the hero of a romance winning a particular woman's love may be worth more than are all the other things he covets put together. He may gamble all else for it. But it has other expressions too, not always tagged “romanticism.” What else but this is “Give me liberty or give me death”? In the great religions of the world the same attitude may be found, though it has never been so described:

TI Matthew 13: 44–6: Again, the kingdom of heaven is like unto treasure hid in a field, the which when a man hath found, he hideth, and for joy thereof goeth and selleth all that he hath and buyeth that field. Again, the kingdom of heaven is like unto a merchant man, seeking goodly pearls, who, when he had found one pearl of great price, went and sold all that he had, and bought it.

What answers to romanticism in Greek antiquity is the heroic code.

“Irony,” says Quintilian, is that figure of speech or trope “in which something contrary to what is said is to be understood” (contrarium ei quod dicitur intelligendum est). His formula has stood the test of time. It passes intact in Dr Johnson's dictionary (“mode of speech in which the meaning is contrary to the words” [1755]), and survives virtually intact in ours: “Irony is the use of words to express something other than, and especially the opposite of, [their] literal meaning” (Webster's). Here is an example, as simple and banal as I can make it: a British visitor, landing in Los Angeles in the midst of a downpour, is heard to remark, “What fine weather you are having here.” The weather is foul, he calls it “fine,” and has no trouble making himself understood to mean the contrary of what he says.

Why should we want to put such twists on words, making them mean something so different from their “literal” – i.e. their established, commonly understood – sense that it could even be its opposite? For one thing, humour. For another, mockery. Or, perhaps both at once, as when Mae West explains why she is declining President Gerald Ford's invitation to a state dinner at the White House: “It's an awful long way to go for just one meal.”

At some time in the course of his life Plato acquired such thorough knowledge of mathematics that he was able to associate in the Academy on easy terms with the finest mathematicians of his time, sharing and abetting their enthusiasm for their work. The Academicorum Philosophorum Index Herculanensis (ed. Mekkler, p. 17) goes so far as to picture Plato as “masterminding” (ἀρχιτεκτον) the researches of his mathematical colleagues. This we may discount as eulogistic blow-up. Not so its further statement that Plato “set problems” to the mathematicians. Elsewhere I have argued for the credibility of the well-known report in Simplicius:

There is no good reason to doubt that Plato had been the first to project the idea that the apparently inconstant motions of the planets could be accounted for by compositions of invariantly constant circular motions proceeding in different planes, directions, and angular velocities. If Plato could hit on this powerful and fertile notion which “under the name of the Platonic axiom was to dominate theoretical astronomy for twenty centuries” and could propound it in a form which would strike Eudoxus and other practicing mathematical astronomers not as a pretty fancy but as a workable hypothesis, he must have been accepted by them as no dabbler in their business but as a student of their subject who understood it so well that his vision of progress in it might even be at certain points ahead of theirs.

At the start of the preceding chapter I staked out the claim that through a “Socrates” in Plato we can come to know the thought of the Socrates of history. By the end of the present chapter the reader should be in a fair position to judge if I am making that claim stick.

Let this much be agreed before I start: much as I need Plato's witness, I could not do the job if I had only his. The most we could learn from his writings is that in different periods of his life he puts into the mouth of Socrates philosophies which are not only different but, in important respects, antithetical. And that of itself would not give a particle of support to my claim. For there is no intrinsic reason why both of these philosophies, despite their polar differences, could not have been Plato's own original creations at different periods of his life. Within the present century we have seen a spectacular illustration of such a shift. In his Tractatus Wittgenstein produced a startlingly original philosophy and published it with the confidence that it would be his last word – indeed the last word – on the subject, announcing in the Preface: “The truth of the thoughts communicated in this work seems to me uncontestable and definitive.

A second inquiry by Arnim rivalled Lutoslawski's for the title of the most maligned work in the history of the stylistic method. This seems rather ironical in view of the fact that its express aim was to endow the chronological conclusions previously arrived at by the stylistic method with such a conclusive force that ‘any opposition would be impossible’. As it was, they did not have this force, simply because they were not absolute. In other words, while they showed that it was probable that certain dialogues belonged together by reason of a common possession of particular stylistic features, they did not show that different groupings according to other features were impossible. Although scholars had found, for instance, that a large number of features connected Soph, and Pol. to the Phil., Tim. and Laws and had consequently assumed their temporal proximity, they had never thought of investigating how many features connected these same two dialogues to, say, Hipp. Mi., Phdo and Crat. Yet it was theoretically possible that such an investigation would unearth a larger number of features than in the former case, and the whole chronological order would have to be revised.

This was the doubt which Arnim wished to eliminate. It meant that every dialogue would have to be compared with every other dialogue, a task which would occupy the lifetimes of several investigators, if the material were to be every possible feature of style. However, it did not have to be, since there existed a smaller, yet self-contained body of material to work on in the reply formulae.

Before the true genesis of the stylistic method in Germany at the hands of Dittenberger, an isolated but important observation was made which, like Campbell's, appears to have passed for the most part unnoticed. In his work on rhetoric1 Blass remarked almost incidentally that in striking contrast to his earlier practice Plato eventually adopted Isocrates’ principle of avoiding hiatus.

It is in the Phdr., at least in the dialogue part, that this avoidance of hiatus first becomes noticeable, since scarcely half as many instances occur as, for example, in the Symp. and Rep. At the same time, Blass pointed out, there is no need to put the Phdr. among Plato's last works, since a certain leaning towards Isocrates, who is mentioned in the dialogue, was only to be expected.

In the Laws, however, hiatus is very definitely avoided, so that in book i for example there are only just over ioo instances on 34 pages (Teubner). Next comes the Phil, with hardly on average 2 instances per page. Finally, in Tim. (87 pages) there are about 50, in Crit. only 5, in Soph. (82 pages) 20, and in Pol. (83 pages) 11.

Blass inferred that Soph., Pol., Phil., Tim. and Crit. belong with the Laws to the last period of Plato's literary career. Both his statistics and his deductions were later confirmed in their broad lines by G. Janell (q.v.) who drew up exact figures of hiatus for the majority of the dialogues.

A comparison of Blass's results with Campbell's makes it immediately apparent that they provide mutual support. The two investigators employed entirely different methods, yet arrived at exactly the same conclusion.

Tiemann's articles have already been referred to in the chapter on Ritter, with whose work they are closely connected, the third being a review of his book. The second, relating to certain reply formulae investigated by Ritter, was intended partly to clarify the statistics for individual formulae which Ritter had combined in groups, partly to correct his figures (Table i I.I p. 88). For the reason mentioned previously (p. 56), it is difficult in the case of reply formulae to declare categorically that one investigator's data are correct, another's incorrect, but Tiemann appears to have been more consistent than Ritter in his specifications.

There are, Tiemann said, three things worth noting in the statistics.

1. (a) Formulae with ὀρθῶς are very rare in the first period, only ὀρθῶς λέγεις being found, 5 times in all. The same formulae are more frequent in the second period, still more so in the third.

2. (b) Almost only the formulae with λέγεις are used in the first period. In the dialogues of the other two periods the percentage of formulae without λέγεις in relation to the total is as follows: Theaet. 58, Phdr. 20, Rep. 74, Soph. 76, Pol. 87, Phil. 65, Laws 41.

3. The small percentage in the Phdr.> Tiemann explained, is of no significance, since the total of 5 formulae hardly permits any comparison. On the other hand the considerable decrease in the Laws is very remarkable, and we should perhaps recognise it as a sign of increasing prolixity of expression.

4. […]