| Overview
of Great Vowel Shift
Evidence for GVS from Modern English
Evidence for GVS from Middle English
1. Overview
The Great Vowel Shift was a major linguistic rearrangement
which took place in English in the century or two during and after Chaucer’s
lifetime: perhaps 1350-1550 (?). In date, therefore, it seems reasonable
to say that the GVS was perhaps the most important process in the change
from Middle English to Modern English. In simplified terms, the effects
of the GVS are fairly easy to describe: it raised (and in one case fronted)
Middle English long vowels, causing the high long vowels to become diphthongs,
as the mid long vowels became high long vowels and so on. This was a process
limited only to English: contemporary and neighboring languages like French,
German, and Spanish were entirely unaffected. The shift affected words
of both native ancestry and borrowings from French and Latin, and the
many pairs of words in each category which for morphological reasons had
a short-long alteration in Middle English thus have quite radically differing
pronunciations in Modern English (see below).
Evidence for the Great Vowel Shift comes from a variety
of sources:
Modern English spelling and pronunciation (see 2 below)
Middle English spelling, as follows (see 3 below):
---rhyme words
---indications of vowel length (doubled vowels in spelling)
---comparisons with spellings from French and Latin, especially in borrowed
words)
When we examine all of these kinds of data simultaneously,
a fairly clear picture emerges; what I hope to do in the following pages
is to work through the evidence in order to show how we know what we know
about the GVS, and to clarify the details of the big picture. At the same
time, I hope this discussion will indicate just how central the GVS was
for giving Modern English the form it now has.
2. Evidence for the GVS from
Modern English. When we look at the spelling and pronunciation
of modern English words, we can see some striking patterns in the use
of vowels, especially in words that are (more or less) obviously closely
related to one another:
crime: criminal
please: pleasant
grateful: gratitude
abound: abundant
goose: gosling
and so on (there are many more such pairs). In each of
these cases, we can see that in Modern English, the pattern often seems
to involve a lax vowel paired with a tense vowel articulated higher in
the vocal tract, or (in the case of high lax vowels) paired with a tense
diphthong.
[aj]---[I]
[i]---[ ]
[e]---[æ]
[aw]---[ ]
[u]---[a]
Of course, these correspondences should remind us of a key feature of
the Modern English spelling system (perhaps especially for the front vowels),
where the canonical values for the following vowels are:
‘i’ : [aj]/[I]
‘e’ : [i]/[]
‘a’ : [e]/[ æ]
‘u’ : [u]/[ ]
‘o’ : [o]/[a]
As it turns out, these patterns (so familiar from elementary school lessons
about different pronunciations of words with and without the "silent
e"), are really sufficient to hypothesize the probable existence
of a process like the Great Vowel Shift in an earlier stage of English.
If we conceptualize the two diphthongs on this chart as being "on
top of" or "higher than" the high vowels, and if we number
the positions of the ladder we have created like this:
1 [aj]
2 [i]
3 [I]
4 [e]
5 []
6 [æ],
then we can see that the letter ‘i’ corresponds
to both 1 and 3; ‘e’ corresponds to 2 and 5; ‘a’
corresponds to 4 and 6. The simplest explanation for such a set of correspondence
would seem to be that in an earlier stage of our language (but since our
spelling was generally frozen) vowels that were once close enough together
to be written with a single symbol have diverged, so that vowels that
were once probably close on such a ladder are now farther apart, having
moved up or down by one or two positions. Indeed, we might guess that
the earlier stage either had (a) just three vowels (spelled ‘i’
‘e’ and ‘a’), or else (b) three close pairs of
vowels, where each member of a pair was represented in spelling with the
same vowel letter. In either case, if we imagine the front vowels as having
three basic features of height: high, mid, and low, and that three vowel
letters were used in writing to indicate these degrees of height, then
we could understand the modern system as having gone through a process
like this:
'i' ----> [aj]
or ----> [I]
‘e’ ----> [i]
or ----> []
‘a’ ----> [e]
or ----> [æ]
As a hypothesis, this is really not too bad. We could even improve this
argument by suggesting that the three-vowel hypothesis (option a) is not
sufficient: modern minimal pairs like "site" and "sit"
or "seat" and "set" (as well as other pairs of words
like "Christ" and "Christmas") indicate that the feature
conditioning the splitting of each vowel into two was not contextual or
phonological, but was based upon some difference in the quality of the
vowel. (In other words, if "site" and "sit" both went
back to earlier forms with the same vowel, they ought to have come out
just the same in Modern English: they don’t, so the vowels must
have always been different to some degree). Pushing this line of argument
to its limit, we might guess that in the earlier stage of the language,
the letter ‘i’ represented the sounds [I]
and [i], ‘e’ stood for [e] and [],
while ‘a’ stood for [æ] and for something else, perhaps
a tense version of this same vowel.
Through this kind of logic, looking only at modern spelling, pronunciation,
and the relationship between the two, we can see that something like
the Great Vowel Shift must have happened at an earlier stage of English.
Unfortunately, as the problem with [æ] in this analysis suggests,
our final hypothesis here is still not quite right (the even greater problems
with the back vowels--which I have avoided here--confirms that this answer
will not hold up). In order to really understand how the GVS affected
English, we have to look at texts surviving from before the GVS occurred.
3. Evidence for GVS from Middle
English
a. Rhyme Words When we look at Chaucer, we can
see that he uses rhyming words which do not rhyme in Modern English. The
following examples are typical; all are drawn from the Canterbury
Tales:
Chaucer’s rhymes . . . . . Modern Words . . . . . . ModE Vowels
1. heeth, breeth . . . . . . . .heath, breath . . . . . . . . [i]/[]
2. ye, melodye . . . . . . . . eye, melody . . . . . . . . .[aj]/[i]
3. mayde, sayde . . . . . . .maid, said . . . . . . . . . . .[e]/[]
4. post, cost . . . . . . . . . .post, cost . . . . . . . . . . .[o]/[ ]
5. hous, plenteous . . . . . house, plenteous . . . . . .[aw]/[]
6. cam, ram . . . . . . . . . . came, ram . . . . . . . . . .[e]/[æ]
7. goot, hoot . . . . . . . . . goat, hot . . . . . . . . . . . .[o]/[a]
8. two, so . . . . . . . . . . . two, so . . . . . . . . . . . . .[u]/[o]
9. loude, koude . . . . . . . loud, could . . . . . . . . . .[aw]/[U]
10. wyn, Latyn . . . . . . . wine, Latin . . . . . . . . . . [aj]/[I]
It is important to note that many of these pairs in modern English show
the same alteration of vowels identified as typical of modern English
in part 2 above (1, 5, 6, 10). To this degree, it seems that the Middle
English data does help account for the peculiarities of the modern spelling-pronunciation
system. But the many pairs of rhyme words which work somewhat differently
suggest how poorly our analysis in part 2 (based solely on the Modern
English data) will fare when we take a look at the bigger picture.
Occasionally, Chaucer’s rhymes, however, give us the vital piece
of evidence we need in order to understand the Middle English vowel system
and how the GVS comes to affect his rhyme words so that they no longer
rhyme: consider the following rhymes from near the end of the General
Prologue: 'caas'/'Thomas' ('case,' 'Thomas'), 'cheere'/'manere' ('cheer,'
'manner'). Though the rhyme-vowels in the second word of each pair have
now become unstressed and are generally pronounced as a schwa, the double-vowel/single-vowel
spelling variation is crucial. Where Modern English depends upon a tense/lax
distinction, Middle English, as the pairs suggest, depended upon a distinction
of vowel length--in terms of duration. The modern informal terms of "long
vowels" and "short vowels" are thus a reflection of an
earlier stage of the language where vowel length was a distinctive feature.
b. Double-Vowel Spellings. Double-vowel spellings
in Chaucer are central, then, for understanding the short-long distinction.
Unfortunately, Chaucer’s spelling is not uniform enough to do the
whole job. For example, from the list above, the spellings of "heeth"
and "breeth" do nothing to indicate which word has the long
vowel and which the short. Indeed, although all of the word-pairs on the
list are short-long rhymes, none of them are distinguished in Chaucer’s
spelling, which seems to privilege the visual rhyme over phonological
accuracy. Thus, Chaucer will sometimes spell the same word in two ways:
we see both "tolde" and "toold" in Chaucer, for example.
Sometimes, however, his spelling is quite reliable, writing "wood"
and "good" where other Middle English texts have "wod"
and even "god." Also, Chaucer normally doubles only ‘e’
and ‘o’, though he does occasionally double ‘a’.
Other writers may occasionally double the ‘i’; the long ‘u’
is most frequently spelled ‘ou’.
Given that Chaucer’s double-vowel spellings thus seem to correspond
to vowels long in duration, we can conclude that there were pairs of vowels
in Chaucer’s English which were distinguishable only by length:
they were either "short" or "long." These vowels must
have been quite similar, otherwise Chaucer could not use them to rhyme
with one another. And finally, one member of each pair must have "shifted"
to account for the Modern English descendents of these word pairs. Thus
we can conclude that the Middle English vowel system which was in use
during Chaucer’s time must have been something like this (though
even this is still a simplification):
[I:]/[I]
[e:]/[e]
[:]/[]
[u:]/[u]
[o:]/[o]
[:]/[]
[a:]/[a]
As Chaucer’s usage suggests, members of each pair could be considered
to rhyme with one another, perhaps through a sort of "off-rhyme."
The GVS had as its primary effect the alteration of only the long member
of each pair, as follows:
[I:] Æ
[aj:]
[e:] Æ [I:]
[:] Æ
[e:]
[u:] Æ [aw:]
[o:] Æ [u:]
[:]
Æ [o:]
[a:] Æ [e:]
If this were all that had happened, the Great Vowel Shift would be relatively
simple and obvious. Unfortunately, the GVS is not the end of the story.
Immediately after the GVS, for example, we can see that late Middle English
must have had a vowel system something like this:
[I:]/[I]
[e:]/[e]
[]
[u:]/[u]
[o:]/[o]
[]
[a]
[aj:] [aw:]
where the short vowels have not moved and the long vowels have shifted.
We still have certain long-short alternations, where the only apparent
distinctive feature which allowed speakers and listeners to distinguish
certain vowels from one another was vowel length. But these alternations
are no longer connected to words which are morphologically related to
one another (i.e. the vowels in ‘keep’ and ‘kept’
are now [I:] and [],
and are thus easily distinguished by some feature other than length:
in this case height). The GVS, then, eradicated from English all words
in which a difference in vowel length alone functioned to indicate a morhphological
relationship; the consequence seems to have been that there was no longer
any reason to maintain vowel length as a distinctive feature in English.
Certainly, it is clear that vowel length (again, in the sense of duration)
is not a feature of the Modern English vowel system. What seems to have
happened to bring about the modern situation is that the distinction between
long and short vowels (perhaps because it no longer paralleled morphological
relationships) disappeared and was replaced by the distinction between
vowels that Fromkin, Rodman, and Hyams describe with their terms "tense"
and "lax." This second change involved a number of rearrangements
involving both short and long vowels, and these rearrangements led to
our modern vowel system.
The best way to think about this "post-GVS rearrangement" is
probably something like this: as length and shortness ceased to be distinctive
features, the long-short pairs of vowels went through some changes which
served to keep them separate. The long vowels tended to move very slightly
up (and become "tense"), while the short vowels moved slightly
downward and are now called "lax." A few other changes, though
seem to have happened at the same time:
[a] became [æ] in many words (but not in some, like ‘father’)
[u] became [U] in some cases,
and []
in others (‘put’ and ‘sun’ are examples)
[u:] from pre-GVS [o:] became either [u] (‘shoot’) or [U]
(‘book’)
[]
(and sometimes [o]) became [a] (‘hot’)
Significantly, other things sometimes happened more or less as exceptions
to these genreal rules. Thus, what must have been a short vowel for Chaucer
in ‘melodye’ [mlodI ]
turns into a tense vowel in the ModE word ‘melody’ [mldi].
The diagram below outlines the ways in which our modern vowels developed
from Medieval (pre-GVS) vowels. Using charts our understanding of Modern
English), we can reconstruct the Middle English pronunciation of vowels
and hence do things like read Chaucer aloud.
|
