Originally, the original manuscript read “of which we have been speaking / which were poor in heart”. In both manuscripts, after the word speaking, Oliver Cowdery inserted of supralinearly, but with a difference. In 𝓟, Oliver initially wrote “of which have been spoken”. Virtually immediately he supralinearly inserted a subject we and the present participle speaking, having crossed out the past participle spoken. In this manner, Oliver restored the original reading in 𝓞 (“which we have been speaking”) except that when he supralinearly wrote the correct speaking in 𝓟, he also added the preposition of after speaking. In other words, the entire supralinear insertion speaking of was written without any change in the level of ink flow. In 𝓞, on the other hand, the of was supralinearly inserted at some later time, given that the ink flow for that of is very uneven and heavier. This difference suggests that Oliver, after adding the of in 𝓟, consciously decided to also insert the of in 𝓞.
The printed editions have struggled with how to deal with this extra of. There is already an of at the beginning of the relative clause “of which/whom we have been speaking (of)”, so there is really no need for one at the end of that clause. Elsewhere in the text, relative clauses involving the verb speak can have of either at the beginning or end of the relative clause, but there are no instances of such relative clauses having of at both the beginning and end of the clause. The text favors of at the beginning of the clause (for some statistics, see the discussion under 1 Nephi 10:16). Thus the original reading here in Alma 32:4, with of only at the beginning of the relative clause, is the expected reading. More generally, however, the Book of Mormon text allows relative clauses to have the same preposition at both the beginning and the end of the clause. For examples involving the preposition in, see under Alma 23:1.
There is also some evidence that Oliver Cowdery struggled with the need for of in this kind of relative clause. Another example where Oliver may have supralinearly inserted an of in 𝓞 at the same time he inserted the of in 𝓟 is found later in the book of Alma:
In that instance, the evidence argues that the of was necessary. For discussion, see under Alma 40:19.
The 1830 typesetter tried to deal with the problem of the two of ’s in Alma 32:4 by considering the second of as part of the following relative clause; that is, he placed a comma after speaking, thus producing the unexpected and rather strange relative clause “of which were poor in heart”. For the 1837 edition, Joseph Smith changed the which to who, not whom (which suggests that he understood that the of belonged at the end of the preceding relative clause). But he did not alter the punctuation, and the 1837 edition ended up replacing Joseph Smith’s who with a whom and leaving the position of the comma unchanged (thus continuing with basically the same difficult relative clause, “of whom were poor in heart”). This interpretation continues in the LDS text. As David Calabro points out (personal communication), the reading with the of (“of which/whom were poor in heart”) implies that only some of the multitude were poor in heart. The original reading, without the of, implies that this “great multitude” of people that came to Alma were all poor in heart.
The 1908 RLDS edition, by moving the comma from after speaking to after the of and by restoring Joseph Smith’s who, correctly reinterprets the of as belonging to the preceding relative clause (“of whom we have been speaking of, who were poor in heart”). Finally, the 1953 RLDS edition removed the repeated of, thus producing “of whom we have been speaking, who were poor in heart”. In accord with the original text of the Book of Mormon, Oliver Cowdery’s inserted of will be removed from the critical text.
Summary: Remove the intrusive of after the word speaking in Alma 32:4; of course, the original uses of which rather than the edited who and whom will be restored (as is generally the case in the critical text).