Several occurrences of ælf have been excluded from this book. One is a scribal error, as the correction of another Anglo-Saxon scribe confirms: the form ‘se ylfa god’ (putatively ‘the god of the ælfe’) for ‘se sylfa god’ (‘God Himself’) in psalm 59 of the Paris Psalter. Some other examples of ælf, however, stand unaltered in their manuscripts, but have not been considered here because I take them to be hypercorrect forms of words in æl-. This position is worth justifying, and offers some tangential support to my arguments above. Ælfmihtig occurs three times in a short text in Corpus Christi College Cambridge MS 320, folio 117, containing formulas and directions for pastoral use, and dating from around 1000: ‘Gelyfst ðū on god ælfmihtine’; ‘Ic þē bidde & bēode þæt þū gode ælfmihtigum gehȳrsum sȳ’; ‘God ælfmihtig gefultumige ūs’ (‘Believe in God Almighty’; ‘I ask and command that you be obedient to God Almighty’; ‘May God Almighty help us’). Ælmihtig never occurs here. The provenance of this manuscript is unknown, but its language is consistently late West Saxon; there is no other instance of initial /æl-/ in the text for comparison.

Ælfþēod- occurs twice in Brussels, Bibliotheque Royale, MS 1650, but curiously the examples are attributed to different hands (both from about the first quarter of the eleventh century): it would appear that hypercorrection was contagious. Hand A, deriving material from the lost, early Common Recension glossary, glossed peregre (‘as though foreign’) with ‘ælfþēodelice’, for ælþēodelice (‘as though foreign’). The largely indistinguishable hands CD, deriving once more from a lost body of glosses, gloss extern peregrinationis with ‘dre ælfþēodi’, presumably for fremdre ælþēodignysse (‘foreign journey abroad’). The hypercorrect forms may or may not originate with the Brussels scribes themselves; each has a correct counterpart in Oxford, Bodleian MS Digby 146, which is textually related, but the principle of lectio difficilior could be invoked.

The hypercorrection here must relate to the fact that groups of three consonants were liable to lose their middle consonant in West Saxon, which would affect ælf-compounds whose second element began with a consonant. How widespread this was or how profound its effects were in the common lexicon is open to doubt, but it had extensive effects on personal names, where æl- for ælf- is well attested in late Old English.