I have uploaded my working "etymological dictionary", such as it is, to Google docs for all to see.

Link.

Perhaps this might be of some use when establishing the new vocabulary; I found it tremendously useful when writing the etymological survey.

I have included some of the suggestions from paarthurnax, but feel free to make further suggestions. Eventually I might make the file editable by anyone.

~Zul

Perhaps zah 'finite,' zaam 'slave,' and zol 'zombie' share a root implying constriction or limitation (of will or of body).

Thank you! A connection between zaam and zol certainly seems very likely. At first I was hesitant to include zah as well, but a bridge between the meanings seems to appear in *zaal 'rim' (assuming that Keizaal is a more or less direct translation of "Skyrim").

I wonder if the first person pronoun zu'u might be related too?

Zu'u might be a stretch. I don't think we can safely assume keizaal is a direct translation given lok 'sky' seemingly doesn't fit, but zaal could mean something close to 'end,' 'edge,' or 'rim' also given unahzaal 'unending.'

Maybe, and this is a stretch, Keizaal originated from a more literal compound Lokahzaal "sky-end."

That's not an idea I'd seriously entertain but it's an attempt to fit the pieces together.

Yesss... I think you'll agree that we don't need to try and shoe-horn lok into the origin of Keizaal. Kei- *'sky' as a separate element might be related to kest 'storm' or even kein 'war', conjuring up images of a stormy or tempestous sky.

I agree that zaal is most certainly the same element as in unahzaal, meaning 'end', 'edge'.

The lok origin seems a bit more possible if you consider it first appears in the Song of the Dragonborn (which was likely the first piece of Dovahzul ever written), and for the sake of the song, Bethesda needed the name to be two syllables rather than three.

This raises an interesting question as we catalogue the etymology of the language - how much do we take Bethesda's process into consideration? Do we try to determine the etymologies that Bethesda likely came up with, or invent ones as long as they are internally consistent? For example, nahl meaning "living" is likely a mistake on Bethesda's part given the root words were nahlaas and laas. Do we acknowledge this and leave it a mystery, or try to invent a reason to explain laas -> nahlaas -> nahl?

My answer to this is, since we don't have Bethesda's notes in front of us and may never know them, all we have is the language itself. As long as we note that our etymologies are hypothetical, I think we should identify whatever meaningful patterns we can find (even if it's highly unlikely Bethesda would have thought of them), and build a foundation that's consistent with the language as we know it so far.

I agree with you. Our research should be based on the language itself. Otherwise our subject becomes literary criticism, or the gaming equivalent thereof. Not a bad thing, but I think most people on this site is more interested in Dovahzul as a language, not a literary device. If we want to extend the vocabulary and grammar of the language, we need to look at it as an actual language and can't keep second-guessing Bethesda.

Regarding the example nahl, from an external point of view we might acknowledge that this was a likely mistake on Bethesda's part -- but from a linguistic point of view the word certainly isn't unexplainable. Similar formations are not uncommon in natural languages. The first that comes to mind is Swedish bil/German Auto 'car', ultimately from Latin auto-mobile 'self-movable'.

That's a great example. That might imply nahl is part of some colloquial vocabulary, or something so commonly used that it would result in that kind of shortening.

Indeed; and it gives us another tool in our “toolbox” of derivational mechanisms. We might for example speculate that other words with one-consonant suffixes, such as mindok 'know' and mindol 'trick', are shortened forms of words that originally ended in syllabic suffixes or elements (cf mindoraan 'understand').

The word aav 'to join' means specifically 'to join forces or efforts.' It could be involved in the etymologies of lahvraan "to gather/muster" and lahvu "army."

Interesting idea! However, it's difficult to see exactly how the words might be related, given the known derivational patterns. For some reason the l of the root must have been dropped, or an l could have attached itself to the beginning of the latter words (which might happen if, for instance, they commonly follow another word that ends in l). Unless we imagine some older developmental stage where the words were *lahaavraan and *lahaavu, in which case the whole first syllable would be dropped in aav. But without any evidence for any of these processes in other words I don't know how to take this further.

Something to consider is that syllables must end in ah rather than aa, so a construction like laavu must be spelled lahvu unless there is gemination with the v. Similarly, a construction like laavraan would become lahvraan if vraan was entirely contained to the second syllable.

-raan is also seen in other verbs, which makes lahvraan most likely lahv/laav + -raan: ahraan 'to wound,' mindoraan 'to understand', saraan 'to wait'.

Possible root TI indicating direction, to, or towards, found in amativ 'onward/forward', kotin 'into', possibly til 'there', tinvaak 'to speak' (speaking to someone), and maybe even tiid 'time' (in the sense of time flowing forward or toward something).

paarthurnax

Something to consider is that syllables must end in ah rather than aa, so a construction like laavu must be spelled lahvu unless there is gemination with the v. Similarly, a construction like laavraan would become lahvraan if vraan was entirely contained to the second syllable.

-raan is also seen in other verbs, which makes lahvraan most likely lahv/laav + -raan: ahraan 'to wound,' mindoraan 'to understand', saraan 'to wait'.

The distribution of sounds is certainly a subject worthy of further study. The rule of aa becoming ah (and ii, oo, uu > ih, oh, uh?) before syllable boundaries seems very plausible. That would offer an alternative explanation for the variation seen in aar 'servant', aam 'serve', ahmik 'service', pointing to a root AA 'serve' rather than AH. (A root consisting of just a long vowel would upset the current thinking on the root structure, but hey, we're still at the drawing-board.)

However I don't see the need for a root LAA as the origin of lahvu, lahvraan, which are both sufficiently explained by the root LAH. The only reason I see is getting slightly closer to including aav 'join' in the family. And then we still have to account for the inexplicable loss of l.

paarthurnax

Possible root TI indicating direction, to, or towards, found in amativ 'onward/forward', kotin 'into', possibly til 'there', tinvaak 'to speak' (speaking to someone), and maybe even tiid 'time' (in the sense of time flowing forward or toward something).

A root TI is in my list. I had not included amativ among its derivatives; that would leave the unusual element ama to be explained. Perhaps we should consider the possibility of root variation with the vowel on both sides of the consonant?

Adding tiid is an intriguing idea. I'll do that.