Here the earliest extant text, the printer’s manuscript, seems to read contradictorily: “the King did awake his servant before he died insomuch that they did pursue Teancum & slew him”. In the 1840 edition, probably as a result of Joseph Smith’s editing for that edition, the plural they was replaced with the singular he. The RLDS textual tradition followed this emendation until the 1908 RLDS edition restored the they, the reading in 𝓟. In other words, the RLDS text reintroduced the seemingly contradictory reading that has both the singular servant and the plural they. David Calabro points out (personal communication) that one could argue that the they here is simply a more generic reference to the Lamanites and that Teancum was pursued by the servant and other Lamanite soldiers.
The 1981 LDS edition removed the contradiction by emending servant to servants. The main reason for accepting such an emendation is that it is supported by the earlier use of the plural servants to describe Teancum’s assassination of Amalickiah, Ammoron’s brother:
In the Book of Mormon, kings and queens each had servants (see, for instance, the discussion under Alma 22:22–23). Another reason, but less conclusive, is that the odds are considerably greater that the plural s could have been accidentally dropped from servants than he could have been replaced by they, although both are possible. There are about 300 cases in the manuscripts where Oliver Cowdery either omitted or added the plural s, including the following passage involving the word servant where the accidental addition of the s occurred in both 𝓞 and 𝓟 (and with the immediate erasure of the s in both manuscripts):
On the other hand, the number of passages where Oliver Cowdery mixed up they and he in the manuscripts is seven, but each was momentary and immediately replaced by the correct pronoun. Statistically, these factors argue that if there is an error in the text here, the original text in Alma 62:36 most likely read in the plural, with servants and they.
Ultimately, the question here in Alma 62:36 is whether the unspecified they can refer to Lamanite soldiers in general. Such a use of they appears to be unique for the text, which helps explain why the 1840 and 1981 emendations were made in the first place. Elsewhere there are 79 instances of “insomuch that they” (where the they requires some referent), and in each case the referent for the they is easily recoverable from the preceding text. The same finding holds for all instances involving other pronouns that require antecedents, as in “insomuch that he” (17 times) and “insomuch that it” (4 times). There is one case, only two verses later, where they refers to Moroni and his men even though the preceding text specifies only Moroni:
However, the text frequently uses a military leader’s name to stand for that leader and his men (as is common in English); for a list of examples and some discussion, see nearby under Alma 62:25. In such cases, the referent for the plural they remains directly determinable. This is not the case here in Alma 62:36; there is no preceding referent for the they in “but behold the king did awake his servant before he died insomuch that they did pursue Teancum and slew him”. Thus the use of they in Alma 62:36 is unique. The most reasonable solution is that the referent for the they is a plural servants. The earliest extant reading in Alma 62:36 is defective, and the critical text will accept the 1981 emendation of servant to servants as the original reading.
Summary: Accept in Alma 62:36 the 1981 LDS emendation of servant to servants, which is supported by the parallel use of servants in Alma 51:34 (which describes the assassination of Amalickiah, Ammoron’s brother); the plural servants is also supported by the original they in the following clause, “insomuch that they did pursue Teancum and slew him”.