1. Very little evidence for copying:

We classified items into this category in three cases: a) When the word did not have a Turkic etymology and thus did not appear to have been inherited, but when we were unable to find a model for it; these words may therefore actually represent innovations in Sakha. b) When competing opinions have been voiced in the literature as to the copied vs. non-copied status of the word. c) When the word did not have a Turkic etymology, and the putative model words (from a formal point of view) differ so greatly in meaning that it is hard to make a case for copying.

2. Perhaps copied:

We classified items into this category in the following cases: a) When the overlap in meanings between the proposed model and the copy was rather tenuous, requiring convoluted explanations. b) When the formal match was not good, but when copying had been proposed by others. c) Items that could either have been derived in Sakha from copied words (in which case they would have been classified as ‘no evidence for copying’), or that could have been copied as a whole (in which case they would have been classified as ‘probably’ or ‘clearly copied’). d) Items where the phonology indicates that they may have been copied, e.g. words beginning with a s- in Sakha that have Turkic cognates also beginning with an s-. Since Turkic word-initial s- has been lost in Sakha, the retention of s- is indicative of a late copy from some Turkic language.

3. Probably copied:

In this category we classified words that have been suggested as copies and the phonology of which are also indicative of copying, but where the formal and semantic overlap with the model word is not perfect. We also classified words in this category that have been suggested as ancient copies into Proto-Turkic, or that seem to be copied from Ket and Selkup, where nothing is really known about the contact situation.

4. Clearly copied:

Words that are clearly copied with a good formal and semantic match to the model language.

Note: When the only ‘derivation’ that had taken place on copied verbs was the addition of the causative suffix (to make the verb transitive) these were classified as ‘clearly copied’, even though in general derived forms were classified as not having been copied. The reason is that in these cases the causative suffix does not effect a change in meaning (i.e. these are not ‘real’ causatives), only in the syntactic valency of the verb.

In the vocabulary we excluded words that are restricted to individual dialects.

For the morpheme-by-morpheme glosses we hyphenate the actual word. Since Sakha is characterized by bidirectional morphonological alternations of consonants across morpheme boundaries, what appears as the root in our segmentations often is not actually the root form, but the morphophonemically modified form. Thus, it might not always be possible to find the ‘root’ we give in a Sakha dictionary.

It should be noted that the ages given in the database are very approximate! This holds especially for “Proto-Turkic”, for which no observed data exists, and the age of which is debated. For Evenki copies we assume that they took place after the 13th/14th century, but we do not know when Sakha may have stopped copying words from the indigenous languages.