Here at the end of verse 36, we have a number of errors in 𝓟. First of all, Oliver Cowdery initially wrote records as record, but then almost immediately he inserted inline the plural s in the original space between record and the following original of. As described below, this of was later corrected to and (and with somewhat heavier ink flow), but there is no change in the level of ink flow for the inserted plural s. The plural is probably the correct reading since elsewhere in the text we always get the plural form of the word record (that is, records) when it occurs with the word scriptures:
As explained under 1 Nephi 5:21, for each case of record(s) the critical text will follow the earliest textual sources. Thus the plural records is very probably the original reading here in Alma 18:36.
Besides initially writing record in 𝓟, Oliver Cowdery also wrote “of the holy scriptures of the people”. In this instance, Oliver corrected the first of to and (supralinearly writing it as a ampersand) and with somewhat heavier ink flow, which suggests that he made this correction when he proofed 𝓟 against 𝓞. Either and or of will work here. Note, for instance, the phraseology in Alma 18:38 and Alma 37:3 (both listed just above): the first uses and to connect records and scriptures; the second places scriptures in a postmodifying prepositional phrase headed by of. One possible source for Oliver’s initial of here in Alma 18:36 is the of later on in the phrase (“the holy scriptures of the people”). Less plausible in my view is the possibility that Oliver was prompted by the nearby “all the records and scriptures” (in verse 38) to consciously emend the earlier “the records of the holy scriptures” (in verse 36) to “the records and the holy scriptures”. The critical text will accept Oliver’s correction of of to and in Alma 18:36.
Finally, we need to consider the and that comes right before the relative clause “which had been spoken by the prophets”. This and is found in 𝓟 and the 1830 edition, yet elsewhere in the Book of Mormon text whenever and introduces a relative clause, we always have a preceding relative clause conjoined with it. In the earliest text, there are 24 other occurrences of and which, and for each of these there is a preceding relative clause, as in the following example:
There is also one instance of “and also which”, and it too has a preceding relative clause:
The which in eight of the 25 cases refers to people and has thus been edited to who. There are also five instances of and who in the original text, and each of these also has a preceding relative clause (here I exclude, of course, cases where the who is an interrogative pronoun). In other words, the occurrence of and which in the earliest text for Alma 18:36, besides being highly awkward, is completely unsupported by usage elsewhere in the text.
One obvious emendation would be to remove the extra and. The 1837 edition introduced this correction into the text, and it has been retained in all subsequent printed editions. It may seem strange to refer to scriptures as having been spoken, but actually there is another passage later on in the text that supports this reading—in fact, the relative clause is identical to the one here in Alma 18:36:
Further support for the corrected reading in Alma 18:36 (where scriptures are referred to as being spoken) can be found in a revelation given through Joseph Smith to Oliver Cowdery in April 1829, during the time that the Book of Mormon was being translated:
The language is not the same, but it expresses the same idea: namely, that the scriptures contain the spoken word of God.
We might also note that in Alma 18:36 an extra and could have been inserted before the which because of the quite frequent use of and in the preceding text:
Thus accidentally adding an extra and before “which had been spoken by the prophets” (marked above with an arrow) could have readily happened.
Another possible emendation for Alma 18:36 is that during the early transmission of the text, some word or words were omitted between the and and the which. One possibility is that the single word that was accidentally deleted:
One problem with this emendation is that it ends up distinguishing between the holy scriptures and the spoken words of the prophets. Yet those words would have been in some written form, especially here in Alma 18 since the text specifically refers to Ammon relying on the records and scriptures in his teaching of king Lamoni: “and he expounded unto them all the records and scriptures from the time that Lehi left Jerusalem down to the present time” (Alma 18:38). Thus the 1837 reading that removed the and in verse 36 seems more reasonable as an emendation since it specifically claims that the written record (which included the holy scriptures) contained the words of the prophets.
Summary: Accept in Alma 18:36 the 1837 emendation that removed the and before the relative clause “which had been spoken by the prophets”; elsewhere in the Book of Mormon text, a conjoined relative clause is always preceded by another relative clause.