Why does vowel length of Middle English trust not match its probable Old Norse etymon traust?

Question

It's natural to expect retention of vowel length when a word is borrowed from ON to ME, and that's indeed what happened with e. g. ME adj. lọ̄s 'free; loose' from ON lauss.

However, ME n. trust is stated to probably derive from ON traust, which doesn't correspond to that pattern. What might have happened here?

Answer 1

The most probable origin is actually the reflex¹ of an unattested Old English word or a zero-grade² derivation from a Proto-Germanic root. It appears to be phonologically impossible to get it from Norse. OED provides the etymology of trust (n.), the Middle English forms of the word, the additional information about the vowel mismatch from early Scandinavian words and the alternate spellings as below:

Forms:
α. Middle English–1600s truste, Middle English– trust.
β. Middle English–1600s trost, Middle English troste.

Origin: Probably a word inherited from Germanic.
Etymology: Probably the reflex of an unattested Old English *trust (perhaps cognate with Middle High German getrüste company, troop, and the Frankish etymon of post-classical Latin trustis retinue, bodyguard (c800 in the Lex Salica)) < the Germanic base of trust adj.

Alternative etymology.

As with trust v. and trust adj., semantic correspondence with related nouns in North Germanic languages has led to a view that this word is a borrowing from early Scandinavian; however, this presents phonological difficulties and does not easily account for the vowel of the α. forms. For further discussion, see R. Dance in Eng. Stud. 81 (2000) 377.

Form history.

The β. forms probably at least partly reflect association with related nouns in early Scandinavian, although they may also reflect the Middle English spelling convention of writing o for u.

OED mentions another reference in the alternate etymology above by Richard Dance, a professor of Early English in St Catherine's College, Cambridge. I've searched the reference and found an explanation in one of his works where he mentions that it is relatively rare to see a secure evidence for Old English descent of a form first attested in Middle English; but usually dismissed as borrowings from Norse are considered more exciting and favored. He compares this allure of the Scandinavian element to the phenomenon Nornomania³, the alleged obsession with the Scandinavian (‘Norn’) heritage in research on the dialects of Orkney and Shetland. He adds that trust cannot be a loan from Old Norse and explains the origin of trust as below with some additional references:

A classic example is PDE trust, which it is now usually agreed cannot be explained as a loan from ON (cp. OIcel traustr ‘trusty’, treysta ‘to make trusty, trust’) but must be referred to a zero-grade derivation on the same PGmc root, which happens not to be recorded in OE; see e.g. OED s.v. trust adj., d’Ardenne (1961: glossary s.v. trusten), Hoad (1985: 139–40).

_{Dance, Richard, 'Getting a word in: Contact, etymology and English vocabulary in the twelfth century', in Janet Carsten, and Simon Frith (eds), British Academy Lectures 2013-14, British Academy Lectures (London, 2015; online edn, British Academy Scholarship Online, 21 Jan. 2016), https://doi.org/10.5871/bacad/9780197265864.003.0007}

_{Abbreviations in the above citation:
PDE: Present Day English
ON: Old Norse (used to refer to any Scandinavian language variety down to about 1500 AD.)
OIcel: Old Icelandic
PGmc: Proto-Germanic
OE: Old English}

---

_{1. reflex: (linguistics) The descendant of an earlier language element, such as a word or phoneme, in a daughter language. - Wiktionary}

_{2. zero-grade: (Indo-European linguistics) In Proto-Indo-European linguistics, an ablaut form of a root characterized by the absence of the basic ablauting vowel phonemes */e/ and */o/.
*bʰr̥- is the zero-grade of the Indo-European root *bʰer- meaning “to carry, bear”. - Wiktionary}

_{3. Citation from:
Gunnel Melchers | University of Stockholm
https://doi.org/10.1075/cilt.321.11mel}