I have concentrated on those found in the English list for the following reasons:

• It includes many French words
• Many technical words are included in this list whereas they are absent from the shorter French list.
• Technical words tend to be identical or near identical to those in French.

In order to avoid contaminating the output I have had a third person place all the words of 14 letters or more into categories. In doing this they were unaware of which text was the source of the word. I have also generated a random sample of same size from all the words of 14 or more letters contained in the data base and the same person has placed these into their categories.

Many of the words have been placed into more than one category and the analyst was able to create her own categories where she thought it useful.

This same person researched the meaning of difficult words that were included in the output. A comment made by the researcher was that many of these large words in the data-base are of dubious merit, being unnecessary extensions of words where better ones already exist. It is unlikely that any coding would be built on a large proportion of dubious words and should offer scope for narrowing the analysis.

The analyst was my wife Lynn and I am extremely grateful for her help in this aspect and her many useful perspectives on the results I have uncovered.

Of the 340,000 words in the data-base over 36,000 of them are greater than 13 letters in length. And of these 36,000 there are 332 of them found in all the texts (there are also 16 repetitions). This high number came as a surprise to me and highlights how great is the need for a guiding discipline for any person seeking to claim that a text contains anagrams. Difficult and meaningful anagrams can be expected to be found in any lengthy text without their difficulty implying anything about the probability of them being deliberate.

I am interested in the distribution of the words between the different texts and whether they show any indications of being thematic. If so I am interested what the numbers can tell me about the source text. In pursuing this I also have to know whether it is an aberration coming from bias in the word-base.

Stage 1 of this analysis indicated that there was a higher ratio of long words to be found in Nostradamus' text than in that of all the others analysed. This is set out in the following table which also shows the ratio of words in each text that were able to be classified (21% have dubious or nebulous meaning). The ratio expected is 2 to 1 because that is the ratio of the starting positions for anagrams (Nostradamus text 2 times as many letters as the other sources). The actual ratios are over 3.5 to 1.

Long word Anagram Comparison.

This table provides a comparative basis for examining the actual words. I am interested in those groupings of words that exceed expectation (2 to 1) and which are closer to the actual ratio of 3 to 1. There are two tables given below showing the theme and the count for Nostradamus and the Other sources.

Long word Anagram Comparison by Topic- Table 1

Long word Anagram Comparison by Topic- Table 2

The majority of the other categories still show a bias towards Nostradamus text (SEX  7:3, BIOL 13:8, MATHS 7:2, MACHINES 9:3) Some of the topics that don't show any leanings are CHEM 13:10, PHYSICS 3:6, EDUCATION 3:6. 

There are many topics that don't appear at all. These include MUSIC/ POETRY/ ARTS, ASTRONOMY, CLIMATOLOGY, TIME / CALENDARS and LEADERSHIP / KINGSHIP.

It is possible to draw the following conclusions from these tables:

• the data is not evenly distributed but clustered around several themes
• the anagrams from Nostradamus' text form stronger clusters than the other texts

Clustering of words is only a starting point. I am interested in whether they seem associative, i.e. related to each other. I am also interested in whether they are strong or weak words i.e meaningful or nebulous. I therefore present some of the clusters of words found under several of the themes.

The first cluster is that for PROPHECY. This list turns out to have a high degree of association. The words for Nostradamus also have another distinctive character- they actually relate to topics mentioned by Nostradamus. 'Quincentenaries' , an anniversary after 500 years found at VIII.84 Line 3 relates to another line "III.94 line 1 De cinq cens ans plus compte l'on tiendra "    (Five hundred years years more they will keep count of the one)

And the same applies to Tercentenaries. X.100 Line 2 says "Le pempotam des ans plus de trois cens" (The all-powerful one for more than three hundred years). The whole anagram of 'Tercentenaries' is found in 2 places in Nostradamus' text , one is IX.96 Line 1 and  the other is IX.38 Line 4.

The second cluster I present is that of religion. This too shows some highly specific words such as Archdeaconries, Pentacostalism, and Evangelistries, Reincarnations, . It also includes a most distinctive reference that of Tetragrammaton, an aspect of religious naming-code that applies to many religions. Some terms such as unalterableness and marvellousness are much vaguer and clumsy. However to me they seem to form a reasonably powerful set of words for talking about the presence of a God and the spreading of is message.

The other oddity of these two sets of clusters are they emphasise topics that anyone could readily believe might be the themes of any code in Nostradamus' text.

One other cluster that caught my attention was the one on plants for I had not been expecting this theme to be prominent. It seems to be drawing on a description of the growth of a plant or set of plants.

In larger groupings patterning is possibly distributed across several topics. In  Medical terms there seems to a wide  group of illnesses relating to sexuality, digestion, muscle and bone tissue.and nervous disorders.

In each case the contributions from the other texts tend to be less specific and less correlated.

I have presented enough of the lists in full to show that whether it is meaningful or not Nostradamus' text produces clusters that hold central themes and that these in general are related to those one might have anticipated from a 16th Century medical man, interested in herbs, religion and the occult.

From internal analysis it can be seen that Nostradamus' text has a much higher degree of pertinent classification than the Other texts, but I also need to see how this relates to the frequencies in the Word-base. The data needed for this is presented in the table below. It  comes from categorising a randomly selected set of words of about the same quantity as in the foregoing. My generator chose 393 words from the base of 3600. These generated over 470 categorizations.

I have shown the count in the column labelled RCt and alongside of it I have shown the count for Nostradamus text (NCt), adjusted by a conservative factor of 1.5 to allow for the difference in size of the total words each represents.

Table of frequency for Random words from Word Base compared to Nostradamus' Text

Although this can only be a rough evaluation it shows that there is once more a peculiarity in the result. Those areas where Nostradamus' text is over-represented are PROPHECY, WARS, DISASTER, HISTORY, LANGUAGE and PLANTS. Most of these are exactly the topics which might have expected to find in any coding of Nostradamus.

This method cannot tell us which words are right and which are accidental but it seems to allow the trend to be followed. It implies those themes that might be most valuable for further research.

In believe that the factor of 1,5 used to produce Nostradamus' count should in fact be closer to 2.5  (by using the Other Source texts as a guide to normal expectation). This factor would add RELIGION, BIOLOGY, MEDICINE, MATHEMATICS and POLITICS to the themes of special interest.

It is evident there is a bias in the English word base but this does not invalidate the above conclusions. The emphasis on science-technology is high in the word-base used for testing and low on arts and calendar material. This may be a result of the length of words and at lower levels other topics may become more prominent.