Different authors tend to write with different word densities and thus authorship of disputed texts, especially ancient writings, can be assigned or re-assigned. If the word density of one text differs ‘significantly’ (in the statistical sense) from that of another, the authors may be different. The purpose of Counter is simply to count the number of words and the number of sentences in a file containing plain text, and to report the average number of words in a sentence (word density) and the standard deviation.
Counter provides a test of statistical significance of different results.
There are many disputes about the accuracy and reliability of this comparison procedure through use of word densities. Counter does not seek to resolve these difficulties, but provides a standard, open, level playing field for those using this procedure. It is very necessary to prepare your text properly before using Counter, see Preparing Your Text.
For further discussions of the procedure see D. Sansone, J. Roman Studies, 80 (1990) 174-177 and references cited therein. A sample text in Latin, similar to that used in an analysis by I. Marriott, J. Roman Studies, 69 (1979), 65-77, is included with this package.
It is possible to process up to 100 files at once and to store the results in a file called PRINTOUT.DOC. This file can be printed-out directly or read into a word-processor.
The results of word counts from Counter may differ from those of Blitz Latin. There are three reasons:
- The translator defines sentences in an idiosyncratic manner; for example, a sentence may be deemed to end with a semi-colon.
- Counter does not add Tack-ons (such as -que, as in ‘Romanusque’) as an extra word. That is, Romanusque will be counted as one word and not as two.
- The translator may become over-burdened by excessively complex sentences, when it will give a warning and skip that part of the un-translated text. Counter cannot become over-burdened, provided that it has sufficient computer memory (RAM) in which to operate. The memory requirement is dynamically adjusted according to the size of the file being counted.