Swadesh list is a compilation of concepts for which words are deemed to exist in the largest number of languages. Swadesh list allows to quantify interrelatedness of languages. Swadesh lists are used in lexicostatistics (quantitative assessment of linguistic kinship) and glottochronology (dating of linguistic divergence). Modified Swadesh list, developed by Morris Swadesh in the beginning of the 1950's and modified by S.A.Starostin in the middle of the 1980's, is used as a study tool for the history of the genetically related languages. For glottochronology, which presumes a natural divergence of a homogenous population into geographically separate areas, borrowings are excluded from the list, while in contrast for the ethnolinguistic studies, which assess contribution of differing linguistic families to the modern languages, the fraction of borrowings is the object of the studies. The Swadesh lexicostatistical analysis reflects the admixture effect, not accounted for in the linguistic tree model.

Societal mobility and fluidity of the nomadic peoples brought them in long symbiotic contacts with multitude of other languages, leaving enduring traces of the past events. On a taxonomic level, the ratios of the components roughly reflect demographic ratios, but the societal, temporal, demographic factors, and the process of creolization impact the outcome and blur the picture, imposing a level of uncertainty that combines with the innate inaccuracy of the linguistic analysis. Under most thorough, detailed typological and statistical analysis the margin of error produced by the Swadesh list method is quite compatible with the magnitude of the results; 10%, 20%, and up to 50% error margins can be routinely suspected, and any claim to accuracy to the last significant digit is an empty pretense.

Under prevailing doctrine advocated by V.I. Abaev, Ossetic is a leftover of a language that used to span Eurasia from Pacific to Adriatic, in generic terms called Scythian and Sarmatian. Ossetic is typologically agglutinative language, supposedly distinct from the bulk of the surrounding Caucasus languages by elevated level of the Iranic and Indo-European layers, but with the morphology of the surrounding languages, with agglutinative declension (aka declination, i.e. inflection of nouns, pronouns, and adjectives in categories such as case, number, and gender) and some flexive noun conjugation, with innate to the agglutinative languages blurred categories of adjectives, numerals, participles and postpositions, and prominently no category of the grammatical gender. In the linguistic doctrine composed by V.I. Abaev, Ossetic is classed as Northeastern, Scytho-Sarmatian branch of the Iranian languages. Ascribing Ossetic, with its key linguistic features of phonology, agglutination, morphology, semantics, and syntax not compatible with IE and Iranian language families to the Northeastern branch of the Iranian languages is a linguistic mockery. The only justification for classing Ossetic with the Iranian languages was its linguistic composition. So far, no academic analysis on the linguistic composition of the Ossetic has been performed, no canonic linguistic analysis with the Swadesh List tool has ever been published neither for Iron, nor for Digor, not for the aggregated “Ossetian” languages. An analysis based on the Wikipedia version of the Swadesh List for Ossetic roughly confirms the eyeball assessment done by  V.I. Abaev that Ossetic lexicon is 80% non-IE, that only about 10% of the Ossetic lexicon belongs to the Iranian family.

The Russian authoritative academic encyclopedia “Languages of Russian Federation” states in 2 places the Caucasian substrate of the Ossetian, although in a funny twisted form: “…there is a reason to talk about “Caucasian substrate” of the Ossetian language” (p. 393), “Words from Caucasian languages constitute the main lexicon of the Ossetian language, so we are forced to talk about its Caucasian substrate…” (p. 403).  The canonic definition of a linguistic substrate reads: “An indigenous language that contributes features to the language of an invading people who impose their language on the indigenous population”. The poor Russian academic linguists found strength to subtly state the Caucasian substrate twice, 10 pages apart in a 15 page article, and no strength to explicate what it means: that the base lexicon, vocalism, phonetics, morphology, nouns, pronouns, verbs, grammar voice, participles, conjunctions, grammatical gender, exclamations, grammatical particles, word formation, and syntax have no relation whatsoever to the Iranian languages (“Languages of Russian Federation and neighboring countries”. "Encyclopedia in three volumes, Volume 2, Moscow, Science, 2001, pp. 391-405 ).

The V.I. Abaev's eyeball assessment largely holds true, within the margin of error: unknown association 48.8% vs. Abaev's 60%; Persian 15.3% vs. Abaev's 10%; IE 7.0% vs. Abaev's 10%; and combined Persian + IE 22.3% vs. Abaev's 20%. V.I. Abaev did not specify the ratio breakdown of  Persian + IE, the  10% + 10% composition is implied based on his numerous references, and the combined 20% vs. 22.3% was a pretty good guess: with the assumed margin of error +/- 30%, the statistical association is 22.3% +/- 7%, and the Abaev's number of 20% is squarely within the margin of error.

Abaev's number of 60% ascribed to the Caucasian languages agrees well with the sum of the unknown association 48.8% plus 4.9% Nakh for a total of 53.7% +/- 18%,  it is squarely within the assumed margin of error.

It is apparent that V.I. Abaev grossly underreported the component of the Türkic words in respect to the Persian words, the Türkic words outnumber the Persian words 1.5:1, i.e. 24.0% vs. 15.3%. In fact, the 24.0% number of the Türkic lexemes exceeds the combined number of 22.3% for Persian + IE taken without correction for the margin of error; however accounting for the margin of error would equalize the proportion. In any case, for an objective observer that would apply V.I. Abaev's and Soviet/Russian academia's reasoning for classification, the first preference would be Türkic instead of Iranic, provided that the observer would also completely ignore the 53.7% predominance of the local Caucasian languages. An honest observer would surely not allow such machinations.

Update (7/21/2016) changed association to Türkic (T) in four instances, insignificantly changing the earlier estimates:  Digor + Iron 69 > 77, Digoron (Digorian) 14 > 18, Iron (Iran) 55 > 59; the corresponding percentage of Türkic component has changed 24.0% > 26.8%, 17.1% > 22.0%, 26.8% > 28.8%. Changes occurred at the expense of IE (3 words) and Unknown (1 word) associations (highlighted).