Nearly all Ket words fall into one of two categories: "very little evidence for borrowing", or "clearly borrowed". Nearly all of the latter derive from

Russian. There were a number of instances where a clearly borrowed Russian root was compounded with one or more Native Ket roots, such as qonokssajdo 'to have breakfast' (from Ket qonoks 'morning' + Russian čaj 'tea' + Ket do 'drink'). These are listed as 'clearly borrowed, with only the Russian root discussed in the lower sections of the origins of the loan.

There were only a few instances where words may have come from other Inner Asian languages, notable Turko-Mongol qān 'khan', and possible Iranic na'n 'bread'. There are quite a number of identifiable Selkup (Samoyedic) borrowings, but these mostly involve specialized reindeer vocabulary and most didn't fall into the basic words included in the database. Other than Selkup and Russian, identifiable loans from other languages are few and far between. It is possible that Ket contains loans from other Siberian languages that are completely extinct, but these, unfortunately, are unrecoverable. Much of the very basic vocabulary appears to have clear cognates in Athabaskan-Eyak-Tlingit, suggesting that even many words represented only in Ket are derivative of the Yeniseic proto-language. However, I ignored this in the Age field, given the still hypothetical nature of my "Dene-Yeniseic" work, and so I listed the origin of words without clear cognates in other Yeniseic languages as "no information" rather than "from the proto-language".

Ket is a language that tends to resist borrowing, preferring to coin its own neologisms instead when new items enter the Ket consciousness. One of my favorites is ēɣ suul 'iron snowsled', for 'car'. My Ket informants, who were Russian-Ket bilinguals, tended to reject most obvious Russian words, even for new realia. I think that this Native Ket linguistic conservatism makes the language interesting to include in a study of loanword typology. It also makes it possible to study ancient linguistic relationships, since much of the basic vocabulary, apparently, is original and not borrowed.

The database includes all Southern Ket word forms that are listed in the published dictionaries. Meaning chosen are in most cases exact, or near exact matches for the citation meaning. In cases where the meaning of the entered Ket form is more narrow than that which was asked for, a note is included in the field 'meaning' .

The word form is cited is in standard phonetic transcription, except that I have included my own superscript symbols for the unique prosodic phonemes of Ket: a macron - ā, ē, ī, ō, ū, ə̄, ɯ̄ - denotes high-even tone; an apostrophe after the vowel – a', e', I', o', u', ə', ɯ' - marks a glottalized monosyllable with abrupt rising tone; and a grave accent - - denotes falling tone. Words with two or more vowel moras, including geminate monosyllables, characteristically have a rising falling tone on the first two moras. This features is not marked because it is predictable.

In a number of cases, two or more forms are listed. This includes instances of absolute synonyms (in which case only a single meaning is noted), as well as instances where no general cover term exists, but two or more clear hyponyms exist.

Below are listed the criteria used to determine the analyzability of word forms.

Unanalyzable

This is chosen for the numerous monomophemic roots of Ket, as well as for polymorphemic words that show no possible etymology in the modern language. Infinitive forms historically contained an affix, but in the modern language this affix often has fused to the original root, creating a syllabic tone. When it is not possible to recover the original forms of the morphemes involved, the word is listed as unanalyzable. Loanwords with an inner form that is opaque to native speakers are also routinely listed as unanalyzable.

Semi-analyzable

A. Part of the word appears to be an irregular variant of a root but the remaining portion of the word is semantically opaque or perhaps a cranberry morph.

B. The word appears to contain a fossilized plural or collective suffix.

C. The infinitive consists of elements that are separated in certain finite verb forms by grammatical affixes, yet none of the elements possesses a discrete meaning.

D. Compounds in which at least one root is of questionable etymological meaning.

Analyzable derived

Words derived by suffixation or prefixation. There are relativesly few derivational affixes in Ket.

Analyzable compound

This is by far the largest category of polymorphemic Ket words. In many cases, a root compound also includes the nominalizing suffix /s/. In instances where the root compound does not normally appear without this /s/ I have marked as 'compound', while instances where an independent compound takes /s/ to make a new word I have listed as 'derived'.

Analyzable phrasal

Occasionally, the Ket use a phrase – usually adjectival modifier plus noun – to express the given meaning. I have included those instances where the meaning is expressed by a set phrase, such as qəlaj ūl 'bitter water', for 'coffee'. The prosodic independence of the two elements indicates a phonological phrase rather than a phonological word.

Ket words can be listed as originating from the proto-language in cases were cognates are found in the other extinct Yeniseic languages. In other cases, I have listed 'no information' for Ket words that have no evidence for borrowing. The Russian loanwords fall into two clear groups. "Early Russian loans are those borrowed during contact in pre-revolutionary times (from about 1600 to 1920). These tend to be highly integrated or fairly well integrated both phonologically and grammatically. "Recent Russian loans" are words borrowed during the Soviet period, where the Ket became bilingual. These are usually unintegrated either phonologically or grammatically. I only included cases where the Native Ket speakers actually spontaneously inserted the given words into their speech.