(1) We definitely need this word.
(2) The root is √VṚTṚ. Won't that yield virta? Though estheticly, I like virda more.
(3) We already speak dialects (idiolects, to be precise). It doesn't hurt till we understand one another. It happens in some nomadic tribes like e.g. Khanty, where almost every larger family speaks its own dialect. But it hardens basic schooling, I'm afraid. And anyway, we need somehow approach to a common denominator, if we really want the language to be used.
(4) Today's discussion in the Discord was ugly frustrating. It definitely drives me into a deep depression.

What about an ataque[1] type formation (from some stem meaning "to change, alter")? What would the first vowel be with WIS? Sundomaic /i/ vs. /a/ (of ahosta?)? Does eques imply the former? Know this we must, yess...

The Finnish translation of "variant", muunnelma comes from muu "else", by the way. Which in the very least reminds one that it's good that we know elements like http://eldamo.org/content/words/word-2600596007.html and their proper meaning these days..

[1] A derivation type mentioned in passing in http://folk.uib.no/hnohf/qevolution.rtf
but not dealt with in
http://www.elvish.org/Tengwestie/articles/Renk/noldintenspref.phtml

√VṚTṚ may have re-emerged in later Eldarin as the root √WIS “change”. With the suffix dā (product of an action) we would get *wis-dā > *wizdā. Unfortunately, zd > st in Quenya, which would give vista, conflicting with the word for “air”.

It's possible, though that this word was coined later from virya- “to change, alternate”, giving virda “alteration, variation”.

However, I think some abstract noun formation from virya- would be better, perhaps viryamë or virme?

+Paul Strack When I saw it first, I thought it was strengthening from virya-. Unfortunately, you are right and the stem re-emerges as √WIS. Viryamë is not bad, but virda just sounds powerful.

+Paul Strack fortunately -da isn't the only suffix meaning the result of an action, we have -ta and -na as well which gives further possibilities.

+Ицхак Пензев I like virda as well and would love to find a way to use it, but I can’t figure out how to get there from WIS.

1. Let's assume a LOS/LOR type pair (it may be going too far off course)
2. Let's turn to ahya- and coin ahta "variant".

The LOS/LOR point is what I would have imagined first.
As for ahya, I feel like hy is a single consonant there, not a h+y. But I may be wrong, of course.

WIR is already attested with two other meanings: weave and fresh. I don’t think adding a third is a good idea, since this is no evidence for it. As for ahya, we don’t know it’s root, so I think sticking with WIS is the right path. I still vote for viryamë or virmë for variation.

For dialect, though, we have other options. Maybe a diminutive of lambe: lambinque. Or carpalë = carpa + lé = speech manner.

+Paul Strack I'm fine with *wirme just tried to substantiate my coinage 😊.
Karpale looks too much like a simple verbal noun to me, but karpalenge could work. Or karparaite (cf. S rhaed). Karpalaite "speech-colour" even. I think a bit of a poetic creativity could go a long way.

+Paul Strack Thank you, your arguments are strong. I vote for viryamë. Virmë is too close to √WIR, in my humble and totally subjective opinion.

It's a pity Tolkien never (?) instructed us how these suffixes (like da/na/ta for 'product of an action') would work with derived verbs.

+ܤܡܝ ܦܠܕܢܝܘܤ I find your suggestion intriguing. I have not considered that mode of derivation to date. If not with WIS nevertheless it may yet be very useful in the future.

+Tamas Ferencz thus, what will be our final decision for "variant"?

+Ицхак Пензев well. It would be neat to have a noun which employs a product-suffix da/ta/na (i.e. "a product of change = variant, variation") but I accept that doing that with the WIS root is not satisfactory [1]. I think *viryame is our safest bet.
[1] unless we stick to *virda and say it is a back-formation on analogy of karda "deed". But I don't want to push it :)

+Tamas Ferencz couldn't think of carda. That's great, but perhaps, goes too far. Nevertheless, let's try it. It just sounds!

I think the “product of a change” is a change, not a variant. That’s why a think an abstract formation is better. While I do like the way virda sounds, I think viryamë is probably a better fit.

+Paul Strack ok, let it be so, viryamë. As usually, I do not fight against strong arguments! ;-)