Sound change: the regular,

the unconscious, the mysterious

W. Labov, U. of Pennsylvania

Colloquy on Hermann Paul FRIAS/HPCL May 15, 2009

PowerPoint available on

www.ling.upenn.edu/~labov
 PRINCIPLES
OF THE
Herman Paul
HISTORY OF LANGUAGE
BY

HERMANN PAUL

TRANSLATED FROM THE SECOND EDITION

OF THE ORIGINAL
BY

H. A. STRONG, M. A., LL.D.

NEW YORK
MACMILLAN & CO.
1889
 Summary of the critique of Paul in WLH 1968

• the sole theoretically grounded object of linguistic study is the

idiolect, but there is no explanation as to how community
consensus is achieved.

•change may come about when an individual skews the

distribution of his performance to seek more comfortable behavior
patterns, but this key term is not defined, nor is their any
accounting of the sporadic character of this adjustment.

• the grouping of idiolects with respect to features shows no

organization that would prefer one grouping rather than another.
 Summary of the appreciation of Paul in WLH 1968

Paul’s Principien may be said to reflect the best achievements of

Neogrammarian linguistics.

• maximum rigor of formulation of the regularity principle

• an intensive interest in recurrent regularities

• a concern with phonetic detail

• a feeling for the atypicality of standardized languages among the

totality of languages

• a desire to “portray as many-sidedly as possible the conditions of the

life of language [Sprachleben]”.

• recognition of the dialectological point of view on language change.

regularity
 The Neogrammarian position

Every sound change, inasmuch as it occurs mechanically,

takes place according to laws that admit no exception.
--Ostoff and Brugmann 1878
 Paul on uniformity

We have now to answer the important question, which has

been in recent times the subject of so much dispute: Can
we assert uniformity of sound-laws? . .
Sound-law does not pretend to state what must always
under certain general conditions regularly recur, but
merely express the reign of uniformity within a group of
definite historical phenomena.
PHL 56-57.
 Paul on the regularity of sound change

It must either happen, therefore, that where the same

sound existed previously, the same sound always remains
in the later stages of development as well;
or, where a separation into different sounds has
occurred, there must be a special reason to be assigned;
and further, a reason of a kind affecting sound alone
PHL p. 58
Twentieth century formulation of the Neogrammarian position

Sound-change is merely a change in the speakers’ manner of

producing phonemes and accordingly, affects a phoneme at
every occurrence, regardless of the nature of any particular
linguistic form in which the phoneme happens to occur. . .
The whole assumption can be briefly put into the words:
phonemes change. --Bloomfield 1933:353-4
 Lexical diffusion

The phonetic law does not affect all items at the same time:
some are designed to develop quickly, others remain behind,
some offer strong resistance and succeed in turning back any
effort at transformation. --Gauchat (cited in Dauzat 1922)

We hold that words change their pronunciations by discrete,

perceptual increments (i.e., phonetically abrupt) but severally
at a time (i.e., lexically gradual) --Wang and Chen 1977:150.
 Exemplar theory and lexical diffusion

The assumption that people learn phonetic categories by

remembering many labeled tokens of these categories
explains . . . why leniting historical changes are typically
more advanced for high-frequency words than low-
frequency words.
-- Pierre-Humbert: Exemplar dynamics: word frequency,
lenition and contrast (2000). To appear in J. Bybee and P.
Hopper (eds.), Frequency effects and emergent grammar.
 Resolving the Neogrammarian Controversy
(Labov 1981)

Regular sound change is the result of a gradual transformation

of a single phonetic feature of a phoneme in a continuous
phonetic space.

Lexical diffusion is the result of the abrupt substitution of one

phoneme for another in words that contain that phoneme.
The English Great Vowel shift (Jespersen)

/i:/ [iy] [uw] /u:/

/e:/ /o:/

/æ:/ [ay] [aw] /ɔ:/

 Ogura on lexical diffusion in the English Great Vowel Shift

“The data in Appendix B clearly show that the change of ME i:

does not simultaneously occur but gradually extends its scope
across the lexicon)

We have claimed that the processes of the development of ME i:

and u: have propagated themselves gradually from morpheme to
morpheme.”

--Mieko Ogura 1987. Historical English Phonology: A Lexical

Perspective. Tokyo: Kenkyusha. p. 45)
Phonetic realizations of M.E. u: words in the Survey of English Dialects
Multi-dimensional scaling of all M.E. i: words
 The i:2 class

(i:2): M.E. short i followed by a velar consonant and

/t/ in right, night, fight, sight, etc. In the history of the
best known dialects, the velar was first realized as a
voiceless palatal and then disappeared, with
compensatory lengthening of the vowel.
 The i:3 class

(i:3): long e: followed by g in Old English in lie,

fly, die, and long ɛ: in eye, etc. The /g/ has been
lenited in all the dialects covered by the Orton
Atlas, but the raising of the vowel to i: did not
occur in all dialects.
Muli-dimensional scaling of core M.E. i: words
 Reports of lexical diffusion, 1970-1997
1970
Cheng, Chin-Chuan, and Wang, Wm. S-Y. 1970. Phonological change of Middle Chinese
initials. University of California (Berkeley) Dept. of Linguistics. Project on Linguistic
Analysis, Second Series, 10 CW1 - CW69.
1973
Sherman, D. 1973. Noun-verb stress alternation: an example of the lexical diffusion of
sound change in English. Project on Linguistic Analysis, Reports, Second Series, 17: 46-81.
1976
Barrack, C. M. 1976. Lexical diffusion and the High German consonant shift. Lingua
40:151-75.
Toon, Thomas E. 1976. The variationist analysis of Early Old English manuscript data. In
W. M. Christie Jr. (ed.), Proceedings of the Second International Conference on Historical
Linguistics. Amsterdam: North Holland. Pp. 71-81.
Toon, Thomas E.. 1976. The actuation and implementation of an Old English sound change.
In R. J. Di Pietro & E. L. Blansitt (eds.), The Third Lacus Forum. Pp. 614-622. Columbia,
SC: Hornbeam Press, Inc.
1977
Cheng, Chin-chuan and William S.-Y. Wang. 1977. Tone change in Chaozhou Chinese: a
study of lexical diffusion. In W. S-Y. Wang (ed),The Lexicon in Phonological Change. The
Hague: Mouton Pp. 86-100.
Wang, William S.-Y. and C.-C. Cheng. 1977. Implementation of phonological change: the
Shaungfeng Chinese case. In W. S-Y. Wang (ed.),The lexicon in phonological change. The
Hague: Mouton.
 Reports of lexical diffusion, 1977-1982
1977
Janson, Tore. 1977. Reversed lexical diffusion and lexical split: Loss of -d in Stockholm. In
Wang (ed.), The Lexicon in Phonological Change. The Hague: Mouton. Pp. 252-65.
Lyovin, Anatole. 1977. Sound change, homophony, and lexical diffusion. In W. Wang (ed.),
The Lexicon in Phonological Change. The Hague: Mouton. Pp. 120-32.
1978
Krishnamurti, Bh. 1978. Areal and lexical diffusion of sound change. Language 54. 1-20.
Toon, Tomas E. 1978. Lexical diffusion in Old English. CLS. Papers from the Parasessions
on the Lexicon.
1979
Wang, William S.-Y. 1979. Language change--a lexical perspective. Ann. Rev. Anthropol.
8:353-71.
1980
Milroy, James. 1980. Lexical alternation and the history of English: evidence from an urban
vernacular. In E. Traugott et al. (ed., Papers from the 4th International Conference on
Historical Linguistics. Amsterdam: John Benjamins.
Phillips, B. S. 1980. Lexical diffusion and Southern Tune, Duke, News. American Speech
56:72-78.
1981
Wallace, Rex. 1981. The variable deletion of final s in Latin. Ohio State M.A. Thesis.
Bauer, Robert S. 1982. Cantonese sociolinguistic patterns: correlating social characteristics
of speakers with phonological variables in Hong Kong Cantonese. U. of California Berkeley
dissertation.
 Reports of lexical diffusion, 1982-1987
1982
Li, Paul Jen-Kuei . 1982. Linguistic variations of different age groups in the Atayalic
dialects. The Tsing Hua Journal of Chinese Studies, new series, 14:167-191.
Chan, Marjorie K. M. 1983. Lexical diffusion and two Chinese case studies re-analyzed.
Acta Orientalia 44:117-52.
1983
Phillips, Betty S. 1983. Middle English diphthongization, phonetic analogy, and lexical
diffusion. WORD 34.1: 11-23. April 1983.
1984
Phillips, B. S. 1984. Word frequency and the actuation of sound change. Language 60:320-
42.
Wallace, Rex. 1984. Variable deletion of -s in Latin: Its consequences for Romance. In
Baldi, P. (ed), Papers from the XIIth Linguistic Symposium on Romance Languages.
Philadelphia: J., Benjamins. Pp. 565-577.
1985
Fagan, D. S. 1985. Competing sound change via lexical diffusion in a Portuguese dialect.
Sezione Romanza 27:263-92.,.
1986
Bauer, Robert S. 1986. The microhistory of a sound change in progress in Hong Kong
Cantonese. Journal of Chinese Linguistics 14:1-41.
1987
Lien, Chinfa. 1987. Coexistent tone systems in Chinese dialects. Berkeley: University of
California dissertation.
 Reports of lexical diffusion, 1987-1991
1987
Gamble, G. 1987. Nootkan glottalized resonsants in Nitinat: a case of lexical diffusion. In
W. Wang (ed.), The Lexicon in Phonological Change. The Hague: Mouton. Pp. 266-278.
Ogura, Mieko. 1987. Historical English Phonology: A Lexical Perspective. Tokyo:
Kenkyusha.
1989
Harris, John. 1989. Towards a lexical analysis of sound change in progress. Journal of
Linguistics 25:35-56.
Labov, William. 1989. The exact description of the speech community: short a in
Philadelphia. In R. Fasold & D. Schiffrin (eds.),Language Change and Variation.
Washington, Georgetown U.P. Pp. 1-57.
Phillips, Betty S. 1989. The Diffusion of a Borrowed Sound Change. JENGL 22.2, October
1990
Shen, Zhongwei. 1990. Lexical diffusion: a population perspective and a numerical model.
Journal of Chinese Linguistics 18:159-200.
1991
Ogura, Mieko, William S.-Y. Wang and L. L. Cavalli-Sforza. 1991. The development of ME
i in England: a study in dynamic dialectology. In P. Eckert (ed.), New Ways of Analyzing
Sound Change. New York: Academic Press, pp. 63-106.

.
 Reports of lexical diffusion, 1993-2006
1993
Wang, William S.-Y. and Chinfa Lien 1993. Bidirectional diffusion in sound change. In
Charles Jones (ed.), Historical Linguistics: Problems and Perspectives. London: Longman
Ltd. Pp. 345-400.
1997
Krishnamurti, Bh. 1997. Regularity of sound change through lexical diffusion (A study of s
> h > zero in Gondi dialects. Paper presented to the Panel on Lexical Diffusion at the 16th
Intwernational Congress of Linguists, Paris, July 21.
1998
Krishnamurti, Bh. 1998. Regularity of sound change through lexical diffusion: A study of s
> h > 0 in Gondi dialects. Language Variation and Change 10:193-220.
2006
Phillips, Betty S. 2006. Word frequency and lexical diffusion. New York: Palgrave
Macmillan.
.
 ANAE

Atlas of North
American
EQnglish
Principles of
ANAE
Linguistic
Change, Vol. III
Chapter 13
Words floating on the
surface of sound change
Fronting of /ow/ in North America
Distribution of /ow/ vowels for all of North America. [N=8313].Vowels
before /l/ are shown in black [N=1577].

800

600

400

200

550 1050 1550 2050

F2
Absence of fronting of Vw in vowel system of Alex S., 42,
Providence, RI TS 474.
Fronting of all Vw in the vowel system of Danica L.,
37, Columbus, OH, TS 737.
34 most frequent /ow/ words in the Brown and Telsur corpora
Brown Telsur F2 MEAN F2 SD
no 2201 348 1497 214
home 639 547 1066 176
go 347 626 1386 237
coat 313 43 1302 230
sofa 227 6 1282 168
both 218 730 1202 214
know 179 683 1409 239
most 153 1160 1215 220
old 145 660 1016 175
goal 137 60 1017 110
coke 136 4 1368 191
phone 101 54 1112 191
goat 84 6 1427 243
pole 79 18 932 110
boat 72 165 1293 208
coast 66 61 1321 201
donut 66 1171 161
over 63 1236 1195 200
Polish 59 19 992 135
road 57 197 1327 195
Minnesota 57 13 1282 195
gold 52 60 1009 120
mostly 48 44 1196 207
doe 37 1 1438 238
ago 37 1387 220
fold 31 7 971 182
ocean 27 34 1403 278
cold 26 171 989 143
notice 25 59 1360 277
bowl 23 79 1000 126
low 21 174 1235 128
toast 19 248 1376 219
nose 17 60 1535 147
soda 3 406 1336 182
 100
150
200
250

-250
-200
-150
-100
-50
50
0
Final

Coronal onset

Labial

Following syllables

Labial onset

Nasal

Lateral

/g/ onset

/n/ onset

p <.00001 <.001
/p/ onset

<.05
ocean

nose

coke

boat

home

low

Female

City size
Regression coefficients for the fronting of /ow/ in ANAE data [N=7796]

Age

Formal Style
Surviving regression coefficients in both halves of a random split in the /ow/ tokens [even =
3927, odd = 3869]F1/F2 position of 348 no tokens in /ow/ distribution [N=8296]
250
200
150
100
50
0
no
-50
-100
home
-150
-200
-250

Age
Coronal onset

Formal Style
coke
nose

low
boat

home
ocean
/g/ onset

/p/ onset
Lateral
Final

/n/ onset

City size
Female
Nasal
Labial

Labial onset
Following syllables

p <.00001 <.001 <.05

 Fronting of /ow/ for words before /l/ and others for all of North America and for
the Southeast (South and Midland). Words selected by regression analysis at p
<.001 level as ahead of phonological prediction, light blue; behind, yellow.

1700

1600

1500

1400

All __l
F2 in Hz

1300
SE__l
All
1200
SE

1100

1000

900

800
fold

home

nose

no
Polish

low
old

sofa
cold

gold
pole

over

coke

ocean
coast
boat
bowl

goal
 Paul on the fluctuation of words

Vacillations of pronunciation caused by quicker or

slower, louder or gentler, more careful or more
negligent utterances, will always affect the same
element in the same manner, no matter in what word
it may occur --PHL 59
Absence of fronting of Vw in vowel system of Alex S., 42,
Providence, RI TS 474.
Fronting of all Vw in the vowel system of Danica L.,
37, Columbus, OH, TS 737.
Is home a lexical exception to the fronting of /ow/?

N F1 F2
/ow/ 5950 616 1304
/owl/ 2576 575 1010
home 775 669 1068
Oklahoma 14 589 1045
homebody, etc. 28 641 1037
Omaha 10 655 1119
hoe 26 621 1233
 The /h_m/ effect on the fronting of /ow/

F2
1000 1100 1200 1300 1400
560

owl
580

Oklahoma
600
/ow/
620
F1

hoe
homebody
640
Omaha
660
home
680
the unconscious
Paul on the unconscious character of sound change

. . . there is no such thing as a conscious effort to prevent

sound change. For those who are affected by the chnge
have no suspicion that there is anything to guard against,
and they pass their lives always in the simple belief that
they speak today as they spoke years ago, and that they will
continue to the end to speak in the same way.

PHL 48.
Conscious correction of a completed change: reading
and word lists in New York City
We chased him with a ba--a baseball bat and
yell, “Bad boy! bad, bad! but he was too. . fast,
only my aunt could catch him.

Paul
all
ball
awful
coffee
office
chalk
chocolate
chock
 Project on Cross-Dialectal Comprehension: Gating Experiment 2

Word Phrase Sentence

1. _________ ________________ ___________________________
2. _________ ________________ ___________________________
3. _________ ________________ ___________________________
4. _________ ________________ ___________________________
5. _________ ________________ ___________________________
6. _________ ________________ ___________________________
 The Northern Cities Shift

desk busses

mat head

block
socks boss
Percent correct in Gating Experiments by city and educational
level in Cross Dialectal Comprehension study: block

block living on Senior citizens

one block living on one block

Formant measurements of word lists read by advanced speakers
in Birmingham [B], Chicago[C] and Philadelphia [P]
 Adult change in real time?

Were anyone able to compare the movements which his

organs made in the utterance of a word many years before
with those which he makes at present, he would most
likely find a striking difference. But to make any such real
comparison would be an impossibility.

PHL 48
Real time changes in the lenition of (ch) in Panama City in Cedergren’s
trend study, 1969-1982
Model of generational change of (ch) in Panama City
with no age-grading
 Regression analyses of fronting of (aw) of men and women by decade in
the Philadelphia Neighborhood Study [N=112]

2200
F2 constant + age*F2 age coefficient

2100

2000

WOMEN:
1900 slope = -5.38
r2=.961

1800

1700 MEN:
slope = -6.60
r2=.788
1600
Under 20 20-29 30-39 40-49 50-59 60-
 Lifespan trajectory of a hypothetical sound change for females born in
1962, 1970, 1986 with no adult increment (Labov 1994)

500
1942
450
1946
1950
400
1954
350 b. 1986 1958
1962
300
1966
1970
250
1974
200 b. 1970 1978
1982
150 1986
1990
100 b. 1962
1994
50 1998
2002
0 2006
5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69

Age
 The critical period revised: possible models of adult participation in sound change
120
100
Stable 80
60
Ii  a 40
20
0
5 9 13 17 21 25 19 29 33
Age


120
100
Linear 80
60
Ii  Xi  a 40
20
0
5 9 13 17 21 25 19 29 33
Age


120
100
80
Inverse power 60
a
Ii  Xi  40
20
t2 0
5 9 13 17 21 25 19 29 33
Age
Lifespan trajectory of a hypothetical sound change for females born in 1962, 1970,
1986 with progressively diminishing adult incrementation: cut-off point 17 years

500
1942
450
1946
400 b. 1986 1950
1054
350
1958
300 1962
1966
250
b. 1970 1970
200 1974
r2=.998 1978
150
1982
b. 1962
100 1986
1990
50 1994

0 1998

5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 2002
2006
Age
mysteries
 The individual and the community

all purely psychical reciprocal operation comes to its

fulfillment in the individual mind alone. PHL xxxvii
All that we imagine that we know about the ideas of
another individual depends exclusively upon conclusions
drawn from our own. PHL xxxix
The great resemblance of all linguistic processes n the
most different individuals is the most essential
foundation for an exact scientific knkowledge of these
processes. PHL xlv
 The enigma of uniformity

It is by intercourse, and nothing else, that the language of

the individual is generated. PHL23

If we start with the undeniable truth that each individual

has his or her own language, and that each such language
has its own history, the problem is not so much how from
a language essentially uniform different dialects arise. .
The problem which challenges solution is this: How
comes it that while the language of each individual has its
own special history, this degree of agreement--a certain
greater or less--maintains itself within this
miscellaneously constituted group of individuals? PHL23
Auer and Hinskens (2005): The role of interpersonal
accommodation in a theory of language change

accommodation by the socially integrated speakers of type A.

. . is often surpassed by type D speakers. . . with loose and
ephemeral network contacts who are highly dissatisfied with
their social life. . .
It appears that the best predictor of accommodation is not
frequency of interaction but instead a strong attitudinal
orientation towards the group with whom one wants to
associate -- p. 356
 Auer and Hinskens (2005) conclusion

. . .we certainly cannot exclude the possibility that participants in the

interaction accommodate to each other’s behavior

nor can we exclude the possibility that the frequency of exposure to a

new, spreading feature through intensive network contacts with its
users can lead to the adoption of this variable.
The fronting of (aw) shown by increase of the second formant with
diminishing age in the Philadelphia Neighborhood Study [N=112]
Figure 5. The curvilinear pattern for social class in the fronting of (aw) in
south, down, out, etc. in Philadelphia

20 50
Celeste
20 00 2578 Hz
19 50

19 00

18 50

18 00

17 50

17 00

16 50

16 00

15 50
Lo we r Midd le Uppe r Lo we r Uppe r mid dle Uppe r cl ass
working working working mi ddle cl ass class
class class class
 The communication index C5

Combines answers to questions about the density of

communication on the block:
How many people on the block do you
say hello to?
have coffee with?
ask for advice?. . .
with the proportion of friends who live off the block.
Scattergram of the fronting of (aw) by the communication
index C5 for women in four Philadelphia neighborhoods
 Sociometric position of Celeste S. in the Clark St. network
(Upper figure: advancement of change, lower figure, C5 index).

2368
Matt R.
2079
6 Ginny C.
6.75
2008
Stanley R.
8
2292
Eddie C.
2463 8.75
Celeste S.
2341 10.25
Dot M.
6 2221
Mary J.
6.75
2017 2340
Henry D. Mae D.
8 8.75
 The two-step flow of communication
(Katz and Lazarsfeld, Personal Influence)

1
F
A 2
2 E
C 2
2
B 2

D
 Fronting
Fronting of (aw)
of /ey/ (F2) for 112 speakers
in closed syllables in
in the Philadelphia
made, Neighborhood
pain, lake, etc. Study
by age with partial
regression
by age linesclass
and social for 6 socioeconomic groups in Philadelphia [N=112]
 On the negative effects of sound change

Thus the symmetry of any system of forms meets in sound

change an incessant and aggressive foe. It is hard to realize
how disconnected, confused, and unintelligible language
would gradually become if it had patiently to endure all the
devastations of sound change. Paul 1891
 Natural misunderstanding: cat => cod

Neither my boyfriend Dave nor I are natives to Michigan,

and we are not NCS speakers. Dave had the following
misunderstanding happen three times in the Lansing area, at
two different grocery stores, with two different workers: he
asked for 'catfish' and the man behind the counter gave him
cod, thinking he said 'codfish.’
Natural misunderstanding: pets => pots

Telephone surveyor [Chicago]: Do you have any

pets in the house?
Brian T. [Eastern US] => pots. [thought that 'pot'
was not likely since everyone has pots and pot =
marijuana was too personal; asked for repetition
several times until understood.]
Map 11.8. North American dialects
The U.S. at Night
 The Inland North

U.S. at Night
Milwaukee
Grand
Rapids Syracuse
Chicago Rochester

Flint Buffalo
Detroit
Kenoshat Cleveland
Joliet
Toledo
Settlement patterns 1800-1850 (Kniffen and Glassie 1966)
Figure 3.2. Relationships among America’s Most Populous Metropolitan Areas
 Caption for Figure 3.2.

Relationships among America’s Most Populous Metropolitan Areas. Heavy

lines indicate the largest outflows of interstate telephone calls for a
sample period in 1968. Light lines indicate largest or second largest
outflows of air passengers from SMSAs over 250,000. The flows mapped
from open city symbols are primary flows; flows mapped from solid city
symbols are second largest outflows or traffic shadow assignments if they
are within about 100 miles of their superordinate. With the exception of
places very close to one of the 20 study regions, almost all of the places
east of the Mississippi for which secondary flows are indicated sent their
primary outflow to New York City. Sources: Telephone call data: American
Telephone and Telegraph Company. Air Passenger Data: courtesy of
Professor Michael O. Filani, University of Ibadan.
Figure 9. The UD measure of the Northern Cities Shift
 Summary of the appreciation of Paul in WLH 1968

Paul’s Principien may be said to reflect the best achievements of

Neogrammarian linguistics.

• maximum rigor of formulation of the regularity principle

• an intensive interest in recurrent regularities

• a concern with phonetic detail

• a feeling for the atypicality of standardized languages among the

totality of languages

• a desire to “portray as many-sidedly as possible the conditions of the

life of language [Sprachleben]”.

• recognition of the dialectological point of view on language change.

 Summary of the critique of Paul in WLH 1968

√ the answer must lie in

• the sole theoretically grounded object of transmission among children,
linguistic study is the idiolect, but there is no but there is some evidence for an
ideological substratum among
explanation as to how community consensus adults

is achieved.

•change may come about when an √ principle of maximal

individual skews the distribution of his dispersion combined with
general principles governing
performance to seek more comfortable chain shifting

behavior patterns, but this key term is not √ some progress on the actuation
problem in searching for
defined, nor is their any accounting of the triggering events in both
linguistic and social context
sporadic character of this adjustment.

• the grouping of idiolects with respect to √ ANAE defines dialects by the

active chain shifts in progress, in
features shows no organization that would a phonology organized as
hierarchical sets of subsystems in
prefer one grouping rather than another. the framework of Martinet and
Weinreich