The 60-item Boston Naming Test (BNT) was published in 1983 with norms described as provisional. One recent finding (Thompson & Heaton, 1989) suggests that verbal intelligence, and the Wechsler Adult Intelligence Scale (WA1S)-Revised Vocabulary subtest, in particular, is strongly correlated with BNT performance, and that education is moderately so. High false-positive rates for naming deficit may conceivably result from the application of the published norms with subjects of lower verbal abilities or limited educational backgrounds. To further explore the relationship of naming to other verbal abilities, analyses were undertaken of the correlations between Level 7-9 Gates-MacGinite Reading Vocabulary Test (G-MRVT) and BNT data from 97 schizophrenic, bipolar, and normal subjects. Reading vocabulary is strongly correlated with BNT performance, and the nature of this relationship is essentially the same across the three diagnostic groups. Application of the published norms would have resulted in a high false-positive rate for naming deficit in all groups among subjects with reading vocabularies equivalent to twelfth grade or less. As a word-recognition based reading exercise, the G-MRVT is likely to provide a brain-compromise-resistant index against which the adequacy of naming performances can be assessed. Accordingly, G-MRVT based BNT performance expectation guidelines are presented for use as a complement to the published norms. Other implications are discussed.