3.4 Accounting for the Change 91and adverbs comprise the “intermediate category.” In respect to categoriality,
adjectives would seem to stand to the left, i.e., be more “major,” than adverbs.
Lehmann , for example, asserts that “[i] t is a fact that most of the adverbs in
every language are synchronically derived from nouns, verbs or adjectives”
(2002b: 77), citing the grammaticalization of adverbs from adjectives, as well
as from local nouns, in a number of languages (e.g., ly adverbs in English,
mente adverbs in Vulgar Latin ; l ī ko adverbs in proto- Germanic) but not the
reverse development (2002b: 78).^20
There is thus an important respect in which whilom – in its shift from adverb
to adjective – does not undergo decategorialization. Rather than exhibiting the
expected downgrading of categorial status, it seems to move from a more minor
to a more major word class. As decategorialization is a defi ning characteristic
of grammaticalization, therefore, alternative explanations for the development
of whilom must be sought.
3.4.2 Possible Explanations for the Directionality of Changes in Whilom
3.4.2.1 Degrammaticalization. The most immediate explanation
for the development of adjectival whilom is, of course, degrammaticalization,
which is generally defi ned as movement from more to less grammatical. In the
most complete discussion of degrammaticalization, Norde ( 2009 : 120) defi nes
it as “a composite change whereby a gram in a specifi c context gains in auton-
omy or substance on more than one linguistic level (semantics, morphology,
syntax, or phonology).”^21 Crucial to Norde ’s conception of degrammaticaliza-
tion is that there must be counterdirectionality of the usual cline of grammat-
icalization, thus from infl ectional affi x > clitic > grammatical word > content
word. The phenomenon of degrammaticalization is controversial: It has been
variously understood and defi ned, and its very existence has been denied. Since
only a rather small set of examples of degrammaticalization have been found
and discussed in the literature,^22 and since these would appear to be sporadic
and not generalizable cross- linguistically, the predominant view of most schol-
ars may be summed up as follows:
The evidence is overwhelming that a vast number of known instances of the develop-
ment of grammatical structures involved the development of a lexical item or phrase
20 In discussing the shift from major to minor category, Lehmann ( 2002b : 199) treats nouns,
adjectives, numerals, and verbs as “major” and pronouns, auxiliaries, adpositions, and con-
junctions as “minor.” He is silent about the status of adverbs.
21 For discussions of degrammaticalization, see, for example, Lehmann ( 2002a ), Heine ( 2003b );
Hopper and Traugott ( 2003 : 130– 139).
22 For Heine et al. ( 1991 : 5), examples of degrammaticalization are “statistically insignifi cant.”
Lehmann states even more unequivocally: “No cogent examples of degrammaticalization have
been found” (2002b: 17).