Many linguistic analysis techniques have been met with biased criticism.

A few examples:

- Don Foster: He used word frequencies, punctuation frequencies, sentence lengths, vocabulary analysis, etc. to identify Joe Klein as the author of "Primary Colors". Even after Joe confessed, many people criticized the work as inaccurate. Dr. Foster has had an amazingly high accuracy rate. Yet, every one of his unconfirmed findings (e.g., a previously unknown work by Shakespeare) is still hotly contested.

http://www.amazon.com/exec/obidos/search-handle-url/index=books&field-author-exact=Donald%20W.%20Foster

- The work by Shlomo Argamon, Moshe Koppel, Jonathan Fine, and Anat Rachel Shimoni, "Gender, Genre, and Writing Style in Formal Written Texts" (2003) has been controversial.

- Deborah Tannen has done a significant amount of work on gender differences in vocabulary, yet this work is very controversial.

http://www.amazon.com/exec/obidos/search-handle-url/index=books&field-author-exact=Deborah%20Tannen

(Amazon says that there are 73 books, but that counts duplicates. She may be controversial, but she is prolific.)

- Kevin Leman is well known for his personality profiling based on birth order. Again, this is very controversial work. Yet, based on personalities, he can give a very accurate determination on whether someone is a middle child, has an older male sibling, etc.

http://www.amazon.com/exec/obidos/search-handle-url/index=books&field-author-exact=Kevin%20Leman

- And let us not forget the Myers/Briggs psychological profiling. (This is great for parties but I knew a psych professor who claimed it was all fortune telling.)

- There was even widespread controversy around people who thought the Earth was round. (And in some small groups, there is still controversy here. E.g., "The Flat Earth Society".)

Are these approaches controversial: Sure -- most new things are controversial.

Are they accurate: Within the statistical standard deviation, yes.

I'm not sure if Chaski was addressing any particular analysis method, or

the general analysis approach. For example:

- Gender Determination. The method I used is based on work by Argamon, et al.

- Core words, punctuation, and sentence length. The methods are based on the work of Don Foster.

- Nationality and English competency. Based on work for analyzing competency among foreign language students.

As I mention in the paper, none of these methods are conclusive -- they are profiles and approximations. However, since they measure habitual behavior, they will be consistent for any single author. (Consistent given the accuracy rate.)

[ reply ]

Link to this comment: http://www.securityfocus.com/comments/articles/11419/34140#34140