[W]e empirically assess this claim by formally testing all 469 (unique, US-English) data points from a popular syntax textbook (Adger 2003) using 440 nave participants, two judgment tasks (magnitude estimation and yesno), and three different types of statistical analyses (standard frequentist tests, linear mixed effects models, and Bayes factor analysis). The results suggest that the maximum discrepancy between traditional methods and formal experimental methods is 2%. This suggests that the minimum replication rate of these 469 data points is 98%. (Spouse and Almeida 2012, p. 609, abstract) This can be read as defending either Essentialists consulting of their own intuitions simpliciter, or their selfconsultation of intuitions on uncontroversial textbook cases only. The former is much more controversial than the later.

One might also wonder whether an error rate of 2% really is appropriate for the primary data presented in an elementary textbook. If a geography textbook misidentified 23% of the rivers of the continental United States, or gave incorrect locations for them, or incorrectly reported their lengths, it would forfeit our trust. Analogous claims could be made about any elementary textbook in other fields: an elementary English literature textbook that misidentified the authors of 2% of the books discussed, or their years of publication, etc. Finally, both parties of the debate engage in ad hominem attacks on their opponents. Here is one example of a classic ad hominem or tu quoque attack on Emergentists in defense of constructed examples by Essentialists:

[The charge made concerning armchair data collection] implies that there is something intrinsic to generative grammar that invites partisans of that framework to construct syntactic theories on the evidence of a single persons judgments. Nothing could be farther from the truth. The great bulk of publications in cognitive and functional linguistics follow the same practice. Of course, rhetorically many of the latter decry the use of linguists own intuitions as data. For example, in an important collections [sic] of papers in cognitivefunctional linguistics, only two contributors to the volume present segments of natural discourse, neither filling even a page of text. All of the other contributors employ examples constructed by the linguists themselves. It is quite difficult to find any work in cognitive linguistics (and functional linguists are only slightly better) that uses multiple informants. It seems almost disingenuous to fault generativists for what (for better or worse) is

(Newmeyer 2007: 395)