home
 

The sample size, being the total number of phrases considered in any given theomatic context, determines the total number of occurrences out of which the number of successes will be observed.

Since the author has not, to our knowledge, published an exact set of rules by which he objectively determines what the total sample size will be in any given experiment, but has apparently been somewhat arbitrary in selecting which phrases to use in testing, there is a need to impose a set of rules that will be followed blindly, yet yield samples comparable to the author's.

We have composed the following set of rules, which are approved by the author in T&SM (pp. A-6 to A-8, p.4-42 note 3). We are not aware of any other rules that should be included. These rules are consistently used in our testing.

1. Articles and beginning conjunctions are called "variables."
   There are 16 articles:
   

   Article	Value	Article	Value
   AI	11	THN	358
   H	8	TOUJ	970
   O	70	TO	370
   OI	80	TOIJ	580
   TA	301	TON	420
   TAIS	511	TOU	770
   TAS	501	TW	1100
   THJ	508	TWN	1150

   There are 6 conjunctions:

   Conj	Value
   DE	9
   GAR	104
   KAI	31
   MEN	95
   OTI	380
   OUN*	520

2. Phrases may not end in a variable.
3. Phrases may be made by dropping variables (yet no duplicate sums).
4. Phrases may be formed by removing a beginning conjunction and OUN* if it is preceded by at most one primary (non-variable) word.
5. Chapter and verse divisions are respected.

*Note: OUN (therefore) is an adverb, but for the purposes of phrase construction we treat it as a conjunction.
 

An Observation

One naturally wonders what the impact of such phrase construction rules might be when determining Theomatic factors, and how the randomness of the context might be affected.

The author considers this impact by observing in TII, p. 186, "There are many shorter and repetitive words, particularly the Greek articles (O, TOV, TOU, TW, OI, TOIJ, TOUJ, TWN), that do not use any letters numbered from 1 to 9 in the alphabet. There are slightly more words in the text that are direct multiples of 10 than just one out of 10." This can be verified above, in that 9 of the 16 articles are exact multiples of 10. All 16 are actually multiples of 10 if a cluster radius of 2 is used, as the author most often does.

The author continues, "But these are all jumbled in with words that have odd numerical values. When adding up two or more words for any of the phrase combinations, one will achieve about as many multiples of 10 as would be expected from just random numbers, i.e. one out of every 10 numbers."

This would be true except for rule 3 above, which the author vigorously employs in constructing phrases. This rule will tend to cause factors that are multiples of 10, particularly smaller factors like 20 and 30, to yield unusual significance. Every single time a phrase containing one of these articles happens to be a multiple of 10, there will be at least two such phrases by definition. If there are two such articles in the phrase, the number of such phrases is quadrupled. Non-randomness is evident in this regard in testing: factors that are multiples of 10 are significantly favored by the phrase construction rules. Smaller multiples of 10 are clearly affected in a non-random manner, therefore the factors 10, 20 and 30 are not considered good candidates for Theomatic factors, and are therefore not considered in test results to establish Theomatic significance.  

home
top