I literally wrote in that post "sometimes legions were over/under strength". The point is the standard number for late republican Rome was 5000 men per legion. Continually referring to "6000 men" after this has been pointed out a dozen times is tiresome.
This discussion is out of my usual field, but recently reading Lawrence Keppie, “The making of the Roman Army” he states “Some Roman authors believed that after Marius [...] thus giving the whole legion a strength of 6000.” and “The manpower of the Late Republic is normally assessed by the Roman historians as about 5000 men: many legions were reduced well bellow that figure...” (pp. 64-66).
From these words I can read that Lawrence Keppie although seems to take the 5000 number as the nominal figure, he is cautious, just take a look to the use of the words “normally assessed”.
Adrian Keith Goldsworthy in “The Roman Army at War 100BC-AD200”, page 38, concludes his chapter about the organization of the army stating “When major war of conquest were planned, complete legions were often raised, and the subsequent conflicts fought by large armies, in which a body of 5000-6000 men was a useful sub-unit”. Curious choice of numbers, don’t you think?
My point, even if I am out of my field, is that probably some aggressive words here are not justified, defending a sole number, since there is not total consensus among historians, eventually because the number was not so constant and focusing in a sole number doesn’t allow us to think “out of the box”.
The scholars numbers, right or wrong, are based on facts; like other accounts of people who were there. Revisions of those scholars by modern day commentators is based on nothing but informed speculation. That's not to say numbers from ancient sources are never wrong, I've given many examples where common sense and scholarly consensus agrees they are, but it is deeply problematic to simply revise numbers based on your modern opinion (e.g. "well, it says 30 legions, but they prolly had about 120,000 men"... based on my opinion).
No, you didn't. That was another post. In any case, Polybios 6.21.6-9 would only appear to indicate a minimum (4,000). Hence the view that legions were levied fit for purpose; not specifically to the number 5,000.
The aftual size of the legion varied with time, so when talking about the numbers in a legion depends on which era of Roman history you are talking about. And I suspect that in staffing a legion that you know is going to see a lot of heavy fighting soon, you would want to initially overman the legions to compensate for he inevitable losses that would occur.
Also, the actual legion size could differ from nominal legion size. Losses would not be immediately replaced, and cohorts could be assigned to detached duties during imperial times, so the actual numbers could be less than nominal. But auxiliaris would be assigned to a legion, and although technically not part of the legion, they would effectively increase the manpower of he legion.
As for recruiting, a couple of legions were created by various ways. A couple were created from Roman Navy, the I & II Adiutrix.
Legion sizes certainly varied at times, but what any of that has to do with squaring Duke's wildly contradictory positions is unclear to me. If he wants to explain how his incredulity to Appian and Plutarch, citing 200-450 cohorts awaiting Sulla in Italy on his return, can be squared with his support for Cinna apparently having 300 cohorts when he marched on Rome 3 years earlier, he is welcome to try.