Hungarian Historical Phonology harkály
harkály 'woodpecker; Specht'
First attestation/Old Hungarian data
? 1323 Harkan (place name), see EWUNg
Important dialectal forms
[coming]
Uralic/Ugric/Pre-Hungarian reconstruction
Pre-Hungarian *karV- < ? Proto-Ugric
Abondolo 1996: 80: *karV 'woodpecker'
UEW: PUg *karɜ ‘woodpecker; Specht‘
Status of the Ugric etymology
Improbable (phonologically irregular)
Loan etymology
None suggested so far
Possible substrate loans (new suggestion)
Cognates suggested in earlier research
Khanty: North (O) χăŋra; East (V) kăjərki̮; South (Kr) χăχraj 'Specht' < Proto-Khanty ?
Mansi: North (So) χɔ̄ŋχra, East (KU) χōrχəj, South (TJ) karkāj, West (P) korkəj ‘Specht’ < Proto-Mansi ?
Proto-Ob-Ugric: ?
Commentary
It seems that the reconstruction of a PUg derived form is impossible. The suffixes in Khanty, Mansi and Hungarian are different, and even the unitary PMansi or Hungarian level is difficult to accomplish. The nasal ŋ in some Khanty and Mansi forms remains completely unexplained. Honti notes that the Kaz. form might be somehow influenced by Mansi (but this could hardly explain the nasal ŋ).
It seems highly uncertain whether a Proto-Ugric stem *karV can be reconstructed, as this would not explain the Khanty word-internal j/χ/ŋ. Also, Honti’s Proto-Ob-Ugric reconstruction *kōrγāj and Proto-Khanty reconstruction *kɑ̆rŋaj are problematic in this sense (and Honti’s Proto-Khanty form could not be derived from his Proto-Ob-Ugric reconstruction regularly, as it is hard to understand how POug *γ could result in *ŋ). The velar k in Proto-Mansi *kārkāj would be difficult to derive from POUg *γ. It is clear that the derivational elements in Khanty and Mansi do not correspond to each other regularly.
Abondolo (1996: 80) considers the Ugric words cognates and notes that a metathezis has taken place in Khanty. However, this dos not explain all the discrepancies. Abondolo also notes that the word is a possible derivative, but it might also be considered an original compound involving a participle/agent noun *karV-ja from PU *karV- 'bite' and (in Abondolo's reconstruction) *kïjV 'little bird' (a word attested only in Ob-Ugric). This explanation remains speculative, especially as the reconstruction of a PU verb *karV- is uncertain (also according to UEW) and because the word *kïjV 'little bird' is not attested outside of Khanty and Mansi.
The Ugric words are probably related to each other somehow, and since the reconstruction of a Proto-Ugric form is not possible, we might be dealing with a parallel loanword into Khanty, Mansi and Hungarian. A word for 'woodpecker' could be a substrate loan.
Conclusion
The assumed cognates in Hungarian, Khanty and Mansi cannot be derived from the same reconstructed form regularly. The Ugric words are possible substrate loans.
References
Abondolo 1996: 80: PUg
EWUng 535, s.v. harkály
Honti 1982: 156, no. 324
Riese 2001: 47-48
UEW: Proto-Ugric Uralonet