A Reader in Sociophonetics

The Peripatetic History of Middle English *ѓթ 27

2.3.2 Step 2: Merger of ѓը with ѓթ and convergence with *Ɲ

Ogura also found that by the ¿ rst part of the 15th century, Ɲ FEED, ܭࡃ HEAP,
and ܭࡂ SPEAK all rhymed with each other. She inter prets this to mean that all
three sounds had merged. We agree with Ogura that the change in rhyming
patterns between the late 14th and early 15th centuries is evidence for phono-
logical change. However, we disagree with her as to what that change was. We
suspect that two things had happened. First of all, in Chaucer’s language, ܭࡂ
SPEAK totally merged with ܭࡃ HEAP. Secondly, ܭࡃ HEAP (now including ܭࡂ
SPEAK) moved into a near merger relationship with Ɲ FEED. If the merger
of ܭࡂ SPEAK and ܭࡃ HEAP proceeded by ܭࡃ HEAP moving into the pho-
netic space occupied by ܭࡂ SPEAK, as indicated in Figure 1.3b, then the near
merger of Ɲ FEED and ܭࡃ HEAP would have been a consequence of the ear-
lier near merger of Ɲ FEED and ܭࡂ SPEAK, and the two changes are reduced
to one. We distinguish between the true merger of ܭࡃ HEAP and ܭࡂ SPEAK
and the near merger of Ɲ FEED and ܭࡃ HEAP based on subsequent develop-
ments: ܭࡃ HEAP and ܭࡂ SPEAK shared subsequent developments, while Ɲ
FEED and ܭࡃ HEAP diverged in descendants of Chaucer’s language.

2.3.3 Step 3: Divergence of ѓը/ѓթ from Ɲ and convergence with Ͼ and/or æj

Several pieces of evidence suggest that ܭࡃ HEAP and Ɲ FEED diverged.
First of all, around the turn of the 16th century, the spelling was rein-
troduced for ܭࡃ HEAP, on old Saxon models (Scragg 1974), and this spell-
ing was the norm by the early 17th century. Secondly, according to Dobson
(1968) and others, /i/ in FEED was the norm by the turn of the 16th century.
Third, also star ting in the 16th cent ur y, ܭࡃ HEAP, but not Ɲ FEED, approxi-
mated ۘ NAME and/or *æj DAY, as illustrated in Figure 1.3c. This near
merger continued into the early 18th century in some speech varieties, and,
as we will show (in section 2.3.4) into at least the mid-20th century in some
parts of England. Contemporary evidence for the approximation of HEAP to
NAME and/or DAY, cited by Dobson (1968), Labov (1975, 1994: 298–303),
Harris (1985), and others, indicates that HEAP had the same vowel as NAME
and/or DAY for some speakers in some areas. Most of this evidence is from
non-orthoepical sources. Labov (1994: 302) suggests a socio-economic fac-
tor. The few orthoepists who report a system in which HEAP and NAME
have the same value were the sons of tradesmen, while their contemporaries
who did not report such a system were from landed gentry or noble families.
Further, as noted by Milroy (1992), rejection of evidence in support of sys-
tem (1.3c) (e.g., by Luick 1964: §489, Wolfe 1972, and Cercignani 1981) has
been based in large measure on the presumption that sounds once merged