@ Karl W. Randolph
Thanks for your interesting comments.
(1) You said:
"There’s a big problem with the LXX, namely that by that time many Biblical Hebrew words were forgotten, which resulted in sometimes rather creative translations in the LXX."
Sure, but Hebrew scholars know that the same could be said the other way round, that is, sometimes the MT has 'rather creative' wording unlike a correct translation of LXX (or other ancient versions). This happens because the LXX - more probably - had some base-texts quite different respecting the MT base-texts. I think is unnecessary that I cite you examples to back this known fact (in every case, to you request, I'll able to cite you samples of them). According my viewpoint, it is necessary to get a translation's eclectic mode, collating all the ancient Bible textual witnesses - through the control performed by the textual criticism - try to get the translation more nearer to the 'original' (being respectful to the global Bible context).
(2) You said:
"Nope, neither word has the meaning of “to waver, skid, slip”. חשׁל is used only once in the context of “straggler”. כשׁל is used for tripping or stumbling, leading to a fall. Those two words sounded significantly different in Biblical times. One example which can be shown not to be an example of what you say, is not enough to show that the two graphemes were sometimes swapped in ancient times. You need to show more examples."
The semantic link between the two roots cited by me was foresaw yet in the past. In fact, for a couple of examples (bold is mine), Benjamin Davies (A Compendious and Complete Hebrew and Chaldee Lexicon to the Old Testament
, 1879, p. 236) on חשׁל wrote: “[…] prob
] akin to (which see),כשׁל
[…].”; and, on p. 309 (on כשׁל) he wrote: “[…] perh
] akin to חשׁל
[…] to totter
Along the same lines, Julius Fuerst (A Hebrew & Chaldee Lexicon to the Old Testament
, 1885), on חשׁל wrote: “[…] to reel to and fro, to totter; hence to be week, decaying, feeble, exhausted, a collateral form of כשׁל
(which see) […].” And on p. 705, on כשׁל he wrote: “[…] to totter to and fro; to waver
[…]; not connected with
כסל (which see) but certainly with חשׁל
[…]; hence to stagger
[…], to totter
Another aspect to considerate is linked with the math probability (although I'm not able to quantify it exactly) of the matter. What could be the ratio of probability that two verbal roots like חשׁל/כשׁל which have in common 2 radicals on 3 (66,6 period), along with their meaning in common also, aren't linked each other? I leave this calculation to other more math-experienced than me.
So, taking into consideration the information presented before, it is reasonable to conclude that the letters 'hheth' and 'kaph' were - in ancient - sometimes swapped, even into the Bible Hebrew language. So, it is also reasonable to estimate the possibility that this swapping - between חרב ('sword') and כרב ('cherub') - can be the solution of Job 40:19's meaning.
The last aspect is about the general meaning of the verse.
The traditional translation revolved around the term 'sword' hasn't the full meaning of a wording including the term 'cherub'. According the Bible the major difference between angels and men focusing on power, strenght (Psa 103:20; 2 The 1:7; 2 Pet 2:11). So, if only cherubs were able to draw near a 'Behemoth' (and, 'play with him', according LXX) - like the verse tell us - this give us a clue of what kind of animal it were.
Granted, what has persuaded me (and other scholars in the past) may not persuade others in the same amount, realistically.