Dear Bennett,

Having had Scripta Minoa a couple of months to look at, and encouraged by having your invaluable INDEX, I'm getting to the stage of starting some serious work on Knossos; and I thought I would jot down what my program is, so that we can discuss the common problems. I expect you will already be some way ahead on many of these items. Most of them are to be built up in card-index, and won't be ready to make Notes of for some considerable time.

PRELIMINARIES.

1. Correction of Scripta Minoa drawings to agree with Bennett's readings,
2. Drawing out of Knossos tablets in commodity order.
3. Reverse index of Linear B sign-groups, and allocation of incomplete sign-groups to possible completes.
4. Statistical analysis of syllables and disyllables.

These are mostly for my own information, and I won't do anything in the way of Notes on them, because I am sure you have them well in hand. But if you need help or checking on the last two, let me know. I would prefer not to treat them as high propriety otherwise, but leave them, if necessary, to the time when your own list is ready.

5. Sign-groups common to Knossos and Pylos.
   Divided by probable categories: and into
   A: Identical
   B: Probably inflectional variants
   C: Possibly orthographic variants



6. Analysis of Knossos and Pylos sign-groups of first 2 categories, i.e. personal names. Apparent variations in spelling due to alternative endings applied to the same name-stems.
7. Knossos sign-groups showing inflection or conjunctions (including series common to Pylos).
8. Knossos spelling anomalies, and other evidence for the phonetic grid.
   
   a) Apparent prefixes and compounds.
   b) Letters apparently omitted
   c) Spelling variations of 1 letter (not included in 7, above)
   d) Spelling variations in 2 consecutive letters.

The number of these last 2-letter variations has hitherto been rather disappointing. You get an ending like -ti -ri -ja which one fancies may contain a "dead"-vowel sequence such as t i - r i or the like, but its less inflected form -ti -ra2 turns up with the same first letter.

However, I think the spellings pu2-ra and pu2-ra -- a-ki-ri-jo probably conceal such a piece of vowel-harmony. Something like A - k e - l e - v / A - k i - l i - s ? And the same disyllable seems to behave in the same way in su-ke-re / su-ki-ri-ta though these may be two quite separate names from the same root, and not inflectional or orthographic variations only.

Page 2

Another disyllable (or even trisyllable ??) which you may have noticed is ko-ri-ja-do-no (Knossos) / ko-ri-a2-da-na (Pylos), as used as a description of one of the kinds of the commodity AROM. -do-no and -da-na line up as regards vowel and consonant positions with the places I gave them on the last version of the GRID (Figure 11), for what they're worth. da [not equal to] a2 is a new one on me, and I must look into it a bit more. A connection with a or wa ? The description of the other kind of AROM, incidentally, also presumably recurs in ku-pa-ro (Knossos)/ ku-pa-ro2 (Pylos).

I was rather disappointed to find negative reference to Amnisos as a port of Knossos in the "Palace of Minos"; but I've since been reading Marinatos' account of his 1933- excavations in Praktika, and he seems to have thought that Amnisos was the main port. This doesn't necessarily say much for a-mi-ni-so, of course, but it's worth bearing in mind.

I haven't heard from any of our other colleagues since Scripta Minoa was out; but I expect Ktistopoulos will come out with something before long. It's a pity that some of them, like Sittig, have so far committed themselves to a set of phonetic values that they're bound to be hamstrung by it, in studying the new material.

I'm working on a radio talk I've been asked to give on the background to Myres' book. I've tried to describe the past history of finding writing in the Aegean, and to put over the methods one uses in trying to decipher a script like this. But I don't know if I shall manage to put across just how fascinating it seems to us.

Yours sincerely,

Michael Ventris

I'm beginning to think that I must have left too few vowel spaces, & too many consonants, on the last grid. If for instance, ri ro ro2 re ru ra ra3 ra2 all belong together, or ti to do te de da du, it looks as if there are more than 5 vowels, on the average, or at least a number of homophones.

Taking an average of 61/2 vowels per consonant, even, it only gives you 13 consonant lines: only enough for -, j, w, k, p, t, l, m, n, r, s, z, plus s' or h

This is only what the Cypriot syllabary has. Does the fat of one sign for t, th, d etc go back to the Minoan Script? On the other hand, some of the rarer consonants may have fewer signs, & this may help to redress the balance.