In sum: a mistaken reading in the KJV, ultimately rooted in an ad hoc patch by the Septuagint translator of a corrupted text, appears in Book of Mormon and D & C phraseology.
Reading Isaiah 40:3 I noticed that the Hebrew cantillation marks, which in many cases function as punctuation, indicate a reading that is quite different from the reading tradition that comes via the New Testament from the Septuagint. My translation, accounting for the proper syntax as indicated both by the cantillation marks and by the parallelism inherent in the verse structure:
Symmachus, versum prophetae hebraice scriptum anglice reddens, wrote:A voice [is] crying out: 'In the wilderness, prepare the path of Yahweh; pave [lit. 'make easy, straighten'] in the desert a roadway for our God.
The question is this: does "in the wilderness" describe the voice's location or the location of the path for Yahweh? The thing to note from my translation is that "In the wilderness" does not describe the location of the voice but the location of the path for Yahweh. The Hebrew cantillation makes this clear, as does much stronger evidence that I will discuss below (and I am confirmed in this reading by every modern, scholarly translation out there), but translations like the KJV incorrectly take "in the wilderness" to describe the location of the crying voice. "The voice of one crying in the wilderness" as a result has become a standard tag to refer to a prophet in the empty solitude, but "in the wilderness" does not modify "the voice of one crying."
Now, Isaiah 40:3 does not occur in the Book of Mormon but this phraseology of "one crying in the wilderness" does show up at 1 Nephi 10:8:
Jos. Smith, having been born of destitute parents, wrote:8 Yea, even he should go forth and cry in the wilderness: Prepare ye the way of the Lord, and make his paths straight; for there standeth one among you whom ye know not; and he is mightier than I, whose shoe’s latchet I am not worthy to unloose. And much spake my father concerning this thing.
So, in this vision, Nephi sees John the Baptist in the future and sees him as the voice alluded to in Isaiah 40:3, and moreover he interprets that passage to mean that "in the wilderness" refers to where the voice is, not to where the path should be. This is clearly culled from the language of Matthew 3; the phrase in bold here is paralleled in KJV New Testament at Matt 3:1-3:
An Evangelist wrote:In those days came John the Baptist, preaching in the wilderness of Judaea, And saying, Repent ye: for the kingdom of heaven is at hand. For this is he that was spoken of by the prophet Esaias, saying, The voice of one crying in the wilderness, Prepare ye the way of the Lord, make his paths straight.
The Evangelist is quoting the Septuagint here (φωνὴ βοῶντος ἐν τῇ ἐρήμῳ phonē boōntos en tēi erēmōi), which is clearly not reflected by the Masoretic Hebrew text. The Masoretic text must, however, reflect the original, for reasons I give below. For now, the point to note is that this deviation from the Hebrew passed ultimately into the KJV (probably under the influence of Jerome's Latin Vulgate) and appears in the Book of Mormon and in the D & C as prophetic tag.
What exactly is this deviation? Well, the translation "the voice of one crying in the wilderness" would make the Hebrew participle קוֹרֵ֔א qôrē' into the second member of a construct state and treats "in the wilderness" (בַּמִּדְבָּ֕ר bammidbār) as an adverbial phrase describing where that voice is coming from. Two things make that unlikely to be the case.
The first is the Masorah cantillation/punctuation (unfortunately not showing up here but can be seen by clicking the links above) which demands a strong separation (like a comma) before "crying" and "in the wilderness." The participle is functioning as an adjective, in other words, that modifies "voice," not a noun, and "in the wilderness" does not tell you where the voice is. Thus "a voice crying" rather than "a voice of [someone] crying in the wilderness" is the correct reading, with a strong syntactical pause after "crying". The Masoretes are, admittedly, very late, but that is not the same thing as being wrong or even make them less likely to be right. For the Masoretes, whatever anyone says, were supremely clever textual scholars, as was confirmed by the Biblical material in the Dead Sea Scrolls, which shows that the text they recorded was essentially the same over nearly a thousand years. A thousand years is quite a long time to maintain that sort of textual stability—long, but not unparalleled (e.g. the Vedas were remarkably stable oral productions, which can be proven on historical linguistic grounds by comparing it to the Old Avestan Gathas of Zarathustra). And it's not the cantillation alone that is the issue—it is the consonantal text itself.
An appeal to the masoretic authority is weak in the face of the second point: the parallelism of the verse structure. "in the wilderness" is plainly part of the message of what the voice is crying out, because it is paralleled in the second verse by "in the desert." Thus, everything after "voice crying" is a poetic utterance with two constituent parts to the verse whose inherent parallelism demands that "in the wilderness" be part of the utterance and not modifying the crying voice; it is not an adverbial phrase describing where the voice is but one describing where the path for Yahweh is to be made ready. I show again the Symmachan rendition in order to highlight the parallelism:
Symmachus, versum prophetae hebraice scriptum anglice reddens, iterum wrote:A voice [is] crying out:
'In the wilderness, prepare the path of Yahweh;
pave [lit. 'make easy, straighten'] in the desert a roadway for our God.
The parallelism is clear when accepting the Masoretic punctuation/cantillation, but it is obscured and lost by ignoring the cantillation. And the interlocked parallelisms show us a poet-seer as deeply concerned about the expression of his message as the message itself; we are meant to notice that parallelism, so any version that erases that parallelism is a deviation from poetic structure. The Masoretic cantillation, then, only confirms a poetic structure that is already there.
Notice that there is absolutely no parallelism in the Book of Mormon version; "one crying in the wilderness" is accepted as the whole phrase.
Understanding the impulse of parallelism actually explains the Septuagint reading, which is not a mistake but actually an attempted correction of corrupt text. For if you look at the Septuagint version (and that quoted in Matthew), you see that there still is a parallelism: some version of "make the path ready for god" occurs in both parts of the verse. But notice that in the second part of the Septuagint and Matthean versions, "in the desert" is not there. The translator of the Septuagint was probably working from a corrupted text (the classicists here who do text criticism can see how a parallel phrase could easily be skipped over in transcription) but, as someone who understood Hebrew verse, he still felt the need for parallelism to come out; since the version he had lacked "in the desert" but still had "in the wilderness," he read "in the wilderness" as modifying the location of the voice, not the location of the path for Yahweh. In other words, he removed it from the poetic utterance, since otherwise the parallelism would be lacking in the second part. By removing it from the utterance and making it describe the whereabouts of the "crying voice," he was able to maintain the rest of the parallelism. In the process, however, it deformed the underlying Hebrew text (still maintained by the Masoretes; a corrupted version was in front of the Septuagint translator). That deformation was read into the New Testament quotation as it stands in the KJV. In turn, that deformation appears in the Book of Mormon and D & C.
I have looked also at the Targum Jonathan (an ancient Aramaic translation of the prophetic books for the Hebrew-less peasants), which has similar problems with this verse but fixes them by expanding it so that it too can maintain parallelism. Thus, both the Septuagint and the Targum maintain parallelism.
With the KJV though, it is clearly a mistaken reading, because that version takes "in the wilderness" as an adverbial phrase that modifies "the voice of one crying." It misses the parallelism and in the process creates a kind of epithet for prophets. That epithet shows up in the Book of Mormon and in D & C.
The incorrect reading remains in the KJV, but has been corrected in more modern translations.
In sum, 1) the masoretic punctuation and 2) the inherent structure of the verse demonstrate a solution of the Septuagint translator to a flawed text, and this solution appears as a misreading in the KJV. Blissfully unaware of the textual history of this verse, the LORD put it in 1 Nephi 10:8, as well as D & C 88:66 and 128:20.
One possible apologetic reply is that the prophet Nephi foresaw John the Baptist's KJV version of the Septuagint's version; to the the D & C occurrences, an apologist could suggest, perhaps, that God himself, not knowing anything about Hebrew poetry at all beyond a rudimentary grasp of chiasmus, was influenced by the Septuagint's reading, perhaps unconsciously, chose to bring it out in the KJV, and then inspired Joseph Smith to perpetuate it.
As it stands now, however, we are forced to conclude that Nephi was influenced by the KJV, a scenario which taxes one's sense of chronology (and probability) to the extreme.
Does anyone know if this has ever been noticed? I can't find anything.
EDITED TO ADD: It also looks like the LORD inspired Joseph to perpetuate this misreading in the Inspired Version (JST Matt 3:29; Isa 40:3 left unaltered).