14 November 1986

Dear Tom,


thanks for your engaging letter, which I enjoyed. Why don't you try writing again to Wilf, who is really (I'm sure) a reasonable and agreeable chap too? Aren't we all? (no need to answer that rhetorical question!) All your controversy is actually achieving is to discredit any mathematical approach to attribution studies; which I as a layman feel is a great shame and pity.

   May we, i.e. you and I look again at language(s), which I reckon to know a bit about? Our disciplines have an interface in cryptanalysis, which was part of my wartime trade for nearly four years. I try to keep it up with research into historical cipher and shorthand – all of course at the pencil-and-paper level. Even at that modest stage though one can't help noticing that mathematics has to pay proper attention to everyday experience; thus when Claude Shannon's calculations proved that 15 symbols was the minimum length for unique solution in substitution cipher, every practising solver could have told him it was more like 25.

   And, with great respect, mathematics has also to make sense. In language, -x isn't as acceptable a solution as x, and it's rather unlikely to have a square root. Forgive me, but I continually find myself asking, with the unknown author (unknown in Oxford, anyhow) of 1 Henry VI  'how can these contrarieties agree?'.

   Never mind all that stuff about speaking technically, and wishing to give me the benefit of the doubt. The latter seems to be quite a new (though of course very welcome) development; but let that pass. Let's agree to inhabit a world of discourse where 'cannot be ruled out' actually means what it says: 'cannot be ruled out'.

   OK: then, substituting in the equation, you have said, and you now repeat, that authorship affinity between EI [Edmund Ironside] and Per [Pericles] I/II cannot be ruled out. But you say (do you not?) that the author of Per I/II is the same as the author of Per III-V, namely William Shakespeare of Stratford-upon-Avon in the county of Warwick, gent. So you say, as plainly as possible, that authorship affinity between EI and Shakespeare 'cannot be ruled out'. Not long ago, however, you were saying, loud and clear, in the public prints, which are still on record for everyone to see, that authorship affinity between EI and Shakespeare not only can but has been ruled out, by the very same person using the very same tests and writing in the same language, in the same year.

   I shall seek to devise some tests showing that your style has changed, so dramatically that you must be at least two different people. When the works of 'Merriam' came to be analysed, due credit must be given to your anonymous collaborator!

   Now for your Shakespeare control, who sounds rather like a medium's spirit guide.

   What can this be but an elaborate device for begging the question at issue and guaranteeing wrong results?

    Here again language utterly separates us. You say; assume that your Shakespeare control is incorrect. OK; I'm doing that. Then you say: 'nonetheless the affinity between it and Per I/II remains greater than between Per I/II and EI'. But you don't mean 'nonetheless', do you? You mean 'therefore'. Your subconscious totally refuses to make the assumption you say you're making.  

   Let me try to help! Suppose that Shakespeare wrote early plays in a different early style, exactly as (say) Wagner wrote early music in a different early style. Then the so-called 'Shakespeare control', ie excluding everything he wrote before his 28th year, like the corresponding 'Wagner (etc) control' would automatically show that he could not possibly have written his own early work; it is the perfect instrument for the actual question at issue.

   It seems to me that both you and Wilf make the simplistic and impossible assumption that one artist has one style (however defined) so that two styles means two artists. This is not only baseless and question-begging but utterly contrary to all the known. facts and evidence. Won't you  please either stop it or justify it?

   A sensible Shakespeare control for comparison with early apocrypha would in my way be an analysis of the plays and poems known to be early, by the evidence of the dates on their title-pages; for example VenusLucreceTitus Andronicus; all by 1594, when their author was 30 years old.

   I didn't 'confirm the accuracy' of your Ironside count. In my view, I demonstrated its inaccuracy. Here's another curious contradiction or at least inconsistency. You know as well as I do that Stanley Wells is rather far from infallible. His and Taylor's latest fantasies in the Oxford Shakespeare seem to me a manifest case of folie a deux. I shall be drawing attention, in the public points, before long, to their outrageous mistakes in elementary methodology (all those crazed hallucinations about so-called 'memorial reconstruction' for example). So you'll forgive me if I'm not very impressed by 'Stanley Wells advised me not to count stage directions'! What you ought to be counting, surely, in an authorship context, is authorial usage in a given text. The stage directions in Ironside are certainly authorial; omitting them is just wrong, I think.

   Well, I could rattle on (literally, on this typewriter, which I see has an identity problem with its Is; at least I hope it's the typewriter not me) but perhaps that's enough for the moment. I have some ideas for stylistic testing which I'd like to try out on you some time. How would you like, if you have a spare moment, to compare Ironside, by whatever tests seem suitable to you, with those parts of Edward III which you believe are not by Shakespeare? Your results would be of interest to me; and I have Louis Ule print-outs of those plays which would help me in considering them.

   Of course Edward III, like Pericles is from the same hand throughout. Occam's razor tells you that's the prima facie probability; so does the title-page and other evidence of Tudor drama in general (what proportion of plays of the period can actually be shown to be collaborative). Just stop assuming that two different styles proves two different people (as distinct from the same writer in different stylistic stages or frame of mind), and you'll be, if I may say so, an even more powerful and impressive intellectual force in this field.