pera (Num. 6:5) ('growth') Rashi comments: Its vowel is Patah-(qatan, this is the reading according to Rabbi N.Z.Y. Berlin (Netziv) in Chumash Ha'amek Davar, and also according to Be'er Rehovot; the combined words are Rashi's term for Segol) because it is davuq ('close to'/'in the construct case with') se'ar rosho ('hair of his head'). The Be'er Rehovot queries that Rashi assumes that had the word pera not been in the construct state it would have had some other vowels, and argues that this cannot be so, as nouns of this type (Segolates) do not change in the construct state. He therefore amends the text in Rashi so that the comment applies to the following word se'ar and the word qatan is deleted. Accordingly Rashi is saying: se'ar Its vowel is Patah because it is davuq ('in the construct state') with rosho ('hair of his head'). This interpretation is only partially reproduced by Rabbi H.D. Chavel in his digest of commentaries on Rashi and is difficult to understand there. Rabbi M. Haraz in Leshon Hayyim raises the same question as Be'er Rehovot and writes that possibly there is a scribal error here. He adopts the approach of the Be'er Rehovot in principle, though without discussing the detail.

Amending text is, as is well known, fraught with danger, and tradition has it that a droplet of Rashi's ink ought to be studied for seven clear days. I once saw an explanation of this passage of Rashi, which did not resort to changing the text, but unfortunately have forgotten the reference. Rashi, it said, was bothered that one might read pera with a Qamatz as is done in the pausal form pa:ra; this would leave the following two words se'ar rosho hanging, unconnected with the rest of the pasuq. Therefore Rashi wrote: pera ('growth of') its vowel is Patah qatan (Segol) because it is davuq ('close to') se'ar rosho ('hair of his head'). The word davuq means 'close to,' the ordinary non-technical meaning of the word, rather than 'construct state,' the grammatical meaning. Rashi simply negates the possibility of there being a stop there. This would be a strong break in the flow of the words. Rashi is saying pera flows directly on to se'ar rosho ('the growth of the hair of his head').

Different but consistent

midaber (Num. 7:89) (speaking) What kind of Mem is this? What is the purpose of the Dagesh in the Dalet? In order to answer these two questions Rashi states: Just like mitdaber ('speaking as monologue') it is the glory of the Most High to say midaber ('speaking as monologue') and Moshe hears as a bystander. In the opinion of Rashi the Mem is the present tense Mem of the Hitpael, and the Dagesh in the Dalet indicates the absorption of the Tav of the Hitpael in the Dalet. (Dalet and Tav are phonetically very similar, both letters being dental and plosive, and the absorption of the Tav in the Dalet is therefore likely.) Rashi then goes on to explain the end of the pasuq vaydaber eilav ('He spoke to him') as indicating that Aharon was excluded - only Moshe was addressed.

R' A. ibn Ezra has a different approach. He states: All the grammarians said that midaber is of the Hitpael conjugation and the Tav is absorbed [as Rashi]. In my opinion it is an infinitive like lo yadati daber (Jer. 1:10) and the Mem is an abbreviation of min as is usual in the language. He implies that the Dagesh in the Dalet compensates for the absorbed Nun at the end of min (Nun is a very soft letter and when it requires a silent Sheva it is frequently absorbed in the following letter). Thus according to R' A. ibn Ezra midaber means 'from speaking.' (R' Asher Weiser [editor of R' A. ibn Ezra, Mosad HaRav Kook] adds that this is a form of honor for the speaker.) R' A. ibn Ezra goes on to explain the end of the pasuq vaydaber eilav '[in this manner] He [always] spoke to him'. R' A. ibn Ezra is saying that in that pasuq there is a perat ('detail') first, and then a kelal ('generalization'); daber is Piel in both parts of the pasuq. But Rashi, who interpreted midaber at the beginning of the pasuq as Hitpael, while vaydaber is clearly Piel, then had to interpret vaydaber at the end of the pasuq appropriately.