Hungarian Historical Phonology táltos
táltos 'magician, shaman; magic horse; Zauberer, Schamane; Zauberpferd'
First attestation/Old Hungarian data
1211 Tholtus (see EWUng :146, s.v. táltos)
Important dialectal forms
[coming]
Uralic/Ugric/Pre-Hungarian reconstruction
Pre-Hu *taltVčV or *te̮ltVčV
(Disputed:)
UEW: PUg *tultɜ ‘Zauberei, Zauberkraft‘
WOT: PUg *toltă
Status of the Ugric etymology
Improbable (phonological and semantic problems)
Loan etymology
1. Hu ← West Old Turkic *taltučï ‘the one who exercises a loss of conciousness’ (WOT: 841–846)
2. Hu ← Turkic *tal- ‘schlagen, prügeln‘ (see UEW)
Cognates suggested in earlier research
Khanty: North (Kaz) tǫʌt ‘Hilfe, Linderung (bei einer Krankheit, in der Armut)‘, tǫʌta ‘ohne (große) Mühe, ohne (viel) Lärm (z.B. Beute bekommen); вдруг‘, (N) tolt ‘Riese (eigtl. Zauberer)‘, toltn ~ tolten ‘mit Zauberkraft‘; East (Vj) tolt ‘fever‘ < Proto-Khanty ? *tolt (Honti: *tᴐlt/tolt)
Mansi: North (N) tūlt: tūltėn ‘leicht, einfach’ < Proto-Mansi *tūlt
Proto-Ob-Ugric: (uncertain) *tV̄lt- (Honti 1982: 188, no. 637)
Commentary
The Ugric etymology (originally stemming from Erdélyi 1960) has been debated in recent research literature: Honti (2017: 62–67) provides a good overview of research history and of arguments for and against the etymology, but he does not really address the problematic soundcorrespondences between the Hungarian and Ob-Ugric forms. Abondolo (1996: 44) notes that there are various problems in the Ugric etymology, and it is not even certain that the Khanty words connected here are related.
The change *u > á in Hungarian is irregular: if the etymology goes back to Proto-Ugric/PreHungarian, the vowel á has to reflect earlier *o or *a. It is unclear how Old Hungarian o in the form Tholtus should be read, but the other Old Hungarian forms from the 15th century show a (see EWUng), so o in the 13th century form probably does not reflect [o]. An old (Uralic/PreUgric) *o–a or *a–a stem would yield long *ū in Mansi, so a reconstruction *tola- or *tala could account for the Hungarian and Mansi forms, but the Khanty vocalism is anomalous (not reflecting *o or *a regularly). Honti (1982) reconstructs Proto-Ob-Ugric *tV̄lt, meaning that the quality of the Proto-Ob-Ugric vowel is uncertain; this makes the whole etymology dubious.
The semantic connection of the Hungarian and Ob-Ugric words is dubious. Also the semantic connection between the Ob-Ugric forms is uncertain. It has to be pointed out that the word is also attested in the northern dialects of Khanty and Mansi; Honti (2017) assumes that the inherited Ugric form has been retained in only these dialects, but Erdélyi (1960) assumes the Mansi word is a loan from Khanty. Due to the irregular vowel-correspondences, the borrowing from Khanty to Mansi cannot be ruled out.
Taken together, the problems of phonology and semantics point to the conclusion that the Ugric etymology should be rejected.
The Turkic etymology suggested by WOT seems the most probable option. The suggested West Old Turkic source form *taltučï is ‘the one who exercises a loss of conciousness’ is derived from the Turkic root *tal- ‘to lose strength, to lose consciousness, to fain’. The etymology includes no major problems, but the formation *taltutči has not been attested as such, which makes the loan explanation somewhat hypothetical.
The relationship between the Ob-Ugric forms requires additional research.
Conclusion
The Hungarian word is probably a loan from West Old Turkic loan *taltutčï.
References
Abondolo 1996: 44: ? Proto-Ugric
Erdélyi 1960: Proto-Ugric (Hu + Kh; Mansi borrowed from Khanty)
EWUng: 1475, s.v. táltos: Proto-Ugric
Honti 1982: 188, no. 637, POUg *tV̄lt
Honti 2017: 62–67: Proto-Ugric
Róna-Tas 2017: 56–57: Proto-Ugric or Hu ← Turkic
UEW: Proto-Ugric Uralonet
Vásáry 2012: 37: Proto-Ugric or Hu ← Turkic
WOT: 841–846: Proto-Ugric or Hu ← Turkic