Title: Second Declaration of Robin Lakoff, Ph.D.
Notes: Professor of Linguistics Robin T. Lakoff's statement in which she critiques the FBI's conclusion that the T Docs and U Docs were written by the same individual.




Plaintiff, )
Defendant. )
____________________________ )

I, Robin T. Lakoff, declare as follows:

1. I am employed full-time as a Professor of Linguistics at the University of California at Berkeley. I have held this position since 1976. Previously, between 1972- 1976, I was Associate Professor of Linguistics at the University of California at Berkeley and, from 1969 to 1972, Assistant Professor of Linguistics at the University of Michigan. Currently my responsibilities include teaching courses in the areas of sociolinguistics and pragmatics, and doing research in those areas.

2. I received an A.B. in Classics and Linguistics from Harvard College in 1964; an M.A. in Linguistics and Classics from Indiana University in 1965; and a Ph.D. in Linguistics from Harvard University in 1967.

3. I have published six books and approximately 75 articles, mainly in the areas of sociolinguistics and pragmatics, but also including syntax and semantics.

4. I have reviewed the government's Opposition to Defendant's Motion to Suppress, pages 53-56, and Exhibit 27 of the exhibits submitted with the Opposition in the case of United States v. Kaczynski. Following are my responses:

5. In both the T- and U-documents, there occur a number of words and phrases that, according to the FBI analysis, are sufficiently idiosyncratic as to constitute an unambiguous mark of identity. The defense contends that the stylistic similarities alleged are very much weaker and are not sufficient to establish common authorship.

The government's claim of having established clean identity between the authors of the T- and U-documents is seriously vitiated by the methodology employed in the government analysis and the specific examples cited by the government. To establish the government's claim, it is necessary to show a significant number of distinctive similarities between the two documents that cannot be due to mere coincidence, or to the language common to the two documents being situated in topically similar contexts.

6. Instances (a) where the form found in the two sets of documents is merely less common than an alternative (e,g,, Eat your cake and have it; spellings such as licence and wilfully) and (b) where the form used in both sets of documents is relatively infrequent in colloquial spoken English, but considerably commoner in scholarly written prose such as the Kaczinski essay and the Unabomb manuscript (e, g,, moreover) are not strongly probative as to common authorship. Similarly nonprobative are instances in which the similarities in diction are the nearly-inevitable result of similarities in content. Virtually all of the cases cited in the Search Warrant Affidavit are of these kinds.

7. In order to refute the government's claims, it is not necessary for Kaczynski to show equivalent commonness between alternatives, or a high degree of frequency of occurrence of a form across all discourse genres, all that is necessary is to show that the similarities of usage might easily have arisen for reasons other than common authorship, which has been done in the defense motion and supporting appendices.

Following are point-by-point rejoinders to the claims made by the government in its Opposition (see paragraph 8 below) and in the Declaration of Professor Donald W. Foster, Ph.D., Government Exhibit 27 (see paragraph 9 below). In some cases, where the Opposition itself cites Professor Foster's arguments directly, responses to both documents have been combined in Paragraph 8.

8. Reply to the Government's Opposition to Defendant's Motion to Suppress:

a. "Kaczynski attacks the affidavit's comparative analysis of documents written by the UNABOM suspect and documents known to have been written by Kaczynski. In particular, relying on a declaration from a professor of linguistics who did not read either set of documents..." (Opposition at p. 54:12) and who '"substantially misconstrues or misrepresents the evidence of common authorship' by 'extract[ing] and decontextualizing [sic] particular linguistic similarities.'" Opposition at p. 56:20-22 (quoting Foster Declaration at p. 5, para. 12).

As is apparent from my initial declaration filed in support of the motion to suppress, I did read the Search Warrant Affidavit and its attachments, including the Unabom manuscript and the 1971 essay in their entirety. Although I have not read all of the other T- and U-documents in their totality, doing so would in no way have altered my conclusions and, in fact, would have introduced into the analysis a methodology at variance with that of the government's analyst, so that my conclusions would have become less relevant to that analysis. In other words, to critique any analysis in a valid way, it is necessary to utilize the same body of evidence that was the basis of that analysis.

Agent Fitzgerald's analysis itself was essentially a "decontextualized" set of "extracts" from the texts, made by isolating individual letters, words, or phrases from each text and putting them side by side. Indeed, the Opposition itself acknowledges this fact, stating as follows: "'The analysis performed by Agent Fitzgerald was simply a side-by-side comparison of the content and language in documents written by Kaczynski with documents written by the Unabomber." Opposition at p. 54:19-21. The larger context in which words occurred played no explicit role in Agent Fitzgerald's analysis, which was merely an exercise in counting putatively unusual forms, in isolation. The government nowhere claims that the contexts in which the examples are situated affects their’ interpretation, or strengthens claims of idiosyncratic usage. Therefore, it would have been pointless to have read the documents in their totality, as a total impression would have no bearing on the original FBI analysis of the texts. The defense's analysis essentially reproduced the FBI analyst's method; and therefore any claims that the methodology of the former is superior to that of the latter (cf. Foster Declaration at p. 5, para. 12) are without merit.

b. "Kaczynski's arguments also lose the forest for the trees". Opposition at p. 55:15: (The "forest" here is the general content of the manuscript, the "trees" the specific examples under comparison.)

To the contrary, as noted above, the "trees" are the side- by-side comparisons of specific instances of usage. The point made in the defense critique and in my initial declaration was that the government was trying to claim two points of similarity where only one was real, since the second was wholly derivable from the first. So, there should have been no claims of similarities of phraseology (in particular with respect to nouns, verbs, and adjectives) with reference to passages that had already been cited for similarities of content.

c. You can't eat vour cake and have it.

Opposition, at pp. 53:15-54:7:

It is certainly true, and not contested by Kaczynski, that the commoner form of the phrase is "you can't have your cake and eat it." The defense argument, as noted, is simply that the less common version occurs frequently enough outside either the T- or U-documents that claims of common authorship on this basis are dubious.

9. Professor Foster's Declaration, Exhibit 27 in the Government's Exhibits to Opposition To Defendant's Motion To Suppress:

Many of the points made here are discussed above in Paragraph 8 because portions of Foster's Declaration are cited in the government's Opposition To The Defendant's Motion To Suppress, and thus will only be referred to briefly here.

a. "It is alleged in the Motion to Suppress that 'through a series of false statements, material omissions, and irrelevant information, the Government improperly attempted to suggest that there was evidence that Theodore Kaczynski was the author of the Unabom manuscript'...! find this representation to be unreasonable, unfair, and inaccurate. In my opinion, the analysis by Special Agent Fitzgerald is unassailable in all important respects; conversely, I find the comparative analysis in the Motion to Suppress to be highly unreliable, containing many false and misleading statements and material omissions (Argument, pp. 52-76, 95-99, 109: 'Critique of the FBI Analysis of the T-Documents and the U-Documents,' pp. 1-77) ."

Professor Foster attacks 29 pages of the Motion and the Critique in its entirety, all of which cite specific errors, omissions, or examples of irrelevant information that was included in the Affidavit. I discussed many of these errors in my declaration that was filed in support of the motion to suppress.

Specifically, I found seven categories of errors in the FBI's allegations of identity of authorship: (1) double use of content-linked lexical choices in the T-documents and the U- documents; (2) common similarities in diction, i.e., the use in both sets of documents of the same very common words; (3) the use in both categories of documents of words that, while less common than those cited in (2), nevertheless are of frequent occurrence in scientific, formal, or scholarly prose or are especially apt to be prevalent in writings on topics dealt with by both the essay and the Unabom manuscript; (4) claimed lexical similarities in the T-documents and in the U-documents, in which examination of such claims shows that the words in question are used in different senses, or in different contexts, which renders the superficial similarities meaningless; (5) claimed idiosyncrasies in spelling and/or usage, where the spelling and/or usage, while not the most common, nonetheless is a common secondary form as listed in American dictionaries and manuals of English usage and is the most common form in the United Kingdom; (6) attribution to Mr. Kaczynski as his personal usage of sources translated from Spanish into English, sources translated from English into Spanish, and quotations from other authors or sources; and (7) transcription errors by FBI analysts. In addition, the government failed to acknowledge differences in spelling between the T- and U-documents.

Professor Foster fails to address any of these specific categories of errors or, with one exception, to address or attempt to refute the findings of error enumerated above.[1] Incomprehensibly, he attacks my methodology.

At the same time, without specifically referring to a single example, Professor Foster states that he finds the "comparative analysis in the Motion to Suppress to be highly unreliable containing many false and misleading statements and material omissions." Foster Declaration at p. 4, para. 10. Since he does not reference a single falsity, misleading statement or omission, Foster's opinion deserves no weight.

b. "The evidence of authorial identity rests not on the any one instance of similar thought or language, but in a collocation of shared linguistic habits that extends to spelling, rare diction, grammatical accidence, syntactical habits,...that likewise points to authorial identity of the T- and U- documents." Foster Declaration at p. 3-4, para. 7.

However, if as the defendant claims a great many of these "shared linguistic habits" turn out to be unremarkable for various reasons, then the claimed "collocation" (juxtaposition, across texts) breaks down as evidence of "authorial identity." Professor Foster here is using as presupposition (demonstrated fact) claims that he himself is at the same time merely proffering as opinion.

c. "The evidence of common authorship is far more extensive, detailed, and compelling than the FBI has suggested. The similarities of thought, expression, grammar, syntax, diction, and internal biographical evidence cannot be attributed to mere chance, or to common subject matter, or even to a shared ideology and set of beliefs." Foster Declaration at p. 5, para. 13.

If indeed there is evidence of common authorship that is "more extensive, detailed, and compelling" than the FBI's documents have suggested, it has not been revealed nor was it known to the judge at the time the warrant was issued. Therefore, such evidence, if it exists, is irrelevant.

d. "... but I find that [Lakoff] substantially misconstrues or misrepresents the evidence of common authorship as set forth in the Affidavit. Lakoff extracts and decontextualizes particular linguistic similarities in the T- and U- documents, . ..." Foster Declaration at p. 4-5, para. 12:

No "misrepresentation" or "misconstruction" exists since the Opposition itself (54:19-21) recognizes that the FBI used side-by-side individual comparisons, which the defense mirrored when it used isolated ("decontextualized") extracts to critique the FBI's analysis. Therefore, the defense arguments and Critique explicitly set forth the results of Mr. Kaczynski's review and analysis, in contrast to the government's claim of authorial identity based on errors, omissions and material omissions. Since both I and the FBI use the same methods, there is no basis for Professor Foster's opinion that Agent Fitzgerald's methods and conclusions are "more reliable and tenable" than mine.

10. I declare under penalty of perjury under the laws of the United States that the foregoing is true and

correct to the best of my knowledge and belief. Executed this 14 day of May, 1997, at Berkeley, California.


[1] Professor Foster addresses one Category 5 error, but only to state that "you can't have your cake and eat it too" is the more common version of the phrase; a proposition with which I do not quibble.