Skip to content

Cmudict's dialect empirically determined

Douglas Bagnall edited this page Jul 29, 2012 · 2 revisions

What is the CMU Pronouncing Dictionary's actual dialect? Experiments with grep.

See also CMUdict reads the NZE word list versions 1 and 2.

Calling grep like this:

douglas@kip:~/nze-vox$ grep -w CHEER cmudict.0.7a
CHEER   CH IH R
douglas@kip:~/nze-vox$ grep -w CHAIR cmudict.0.7a
CHAIR   CH EH R

finds the pronunciation of pairs of words. Sometimes a word will have more than one pronunciation, with the first one being preferred.

father-bother merged

FATHER  F AA DH ER
BOTHER  B AA DH ER

cot-caught prefers merged

COT     K AA T
CAUGHT  K AA T
CAUGHT(2)       K AO T

Odd-awed split

ODD     AA D
AWED    AO D

Trap-bath merged

TRAP    T R AE P
BATH    B AE TH

palm-cot merged with occasional alternates

COT     K AA T
LOT     L AA T
LOT(2)  L AO T
PALM    P AA M
PALM(2) P AA L M
POT     P AA T
BOMB    B AA M

Mary-merry-marry merged

MARY    M EH R IY
MARRY   M EH R IY
MERRY   M EH R IY
BURY    B EH R IY
BERRY   B EH R IY
FAIRY   F EH R IY
FERRY   F EH R IY
VARY    V EH R IY
VERY    V EH R IY

Mirror-nearer merged

MIRROR  M IH R ER
NEARER  N IH R ER

Card-cord split

CARD    K AA R D
CORD    K AO R D

Horse-hoarse merged

HOARSE  HH AO R S
HORSE   HH AO R S

Vile-vial split

VILE    V AY L
VIAL    V AY AH L
NILE    N AY L
DENIAL  D IH N AY AH L

Cloth-thought merged

CLOTH   K L AO TH
THOUGHT TH AO T
LOT(2)  L AO T

Law-loss prefers merged

LAW     L AO
LAW(2)  L AA
LOSS    L AO S
LAWS    L AO Z
BALL    B AO L

Here-there split

HERE	HH IY R
THERE	DH EH R

Beer-bear split

BEER	B IH R
BEAR	B EH R
BARE	B EH R

bear-bare merged

BEAR	B EH R
BARE	B EH R

Near-square split

NEAR	N IH R
SQUARE	S K W EH R

Rarely-really split

RARELY	R EH R L IY
REALLY	R IH L IY
REALLY(2)	R IY L IY

Real-reel merged

REAL	R IY L
REEL	R IY L

Dole-dole split

DOLE	D OW L
DOLL	D AA L

Fill-full split

FULL	F UH L
FILL	F IH L

Grown-groan merged

GROWN	G R OW N
GROAN	G R OW N
THROWN	TH R OW N
THRONE	TH R OW N

Yod dropping dropped, occasional alternates

NEW     N UW
NEW(2)  N Y UW
NEW-ZEALAND     N UW Z IY L AH N D
DUKE    D UW K

Beauty-booty split

BEAUTY  B Y UW T IY
BOOTY   B UW T IY
CUTE    K Y UW T
COOT    K UW T

fool-full split

FOOL    F UW L
FULL    F UH L
PULL    P UH L
POOL    P UW L

Pure–poor-pour all split

PURE    P Y UH R
POOR    P UH R
POUR    P AO R

Wine-whine merged, with alternates

WINE    W AY N
WHINE   W AY N
WHINE(2)        HH W AY N

Prints-prince split

PRINTS  P R IH N T S
PRINCE  P R IH N S

Hire-higher merged, with alternates

HIRE    HH AY ER
HIRE(2) HH AY R
HIGHER  HH AY ER

Salary-celery split

CELERY  S EH L ER IY
SALARY  S AE L ER IY

Chair-cheer split

CHEER   CH IH R
CHAIR   CH EH R

Foot-strut split

FOOT    F UH T
STRUT   S T R AH T

Don-dawn split

DON     D AA N
DAWN    D AO N

Knot-naught split

KNOT    N AA T
NAUGHT  N AO T

Bad-lad merged

BAD     B AE D
LAD     L AE D

Pod-pawed split

POD     P AA D
PAWED   P AO D

Have-halve merged

HAVE    HH AE V
HALVE   HH AE V

Alms-arms split

ALMS    AA L M Z
ALMS(2) AA M Z
ARMS    AA R M Z

FARTHER F AA R DH ER
FATHER  F AA DH ER

Pin-pen split

PIN     P IH N
PEN     P EH N

Fill-feel split

FILL    F IH L
FEEL    F IY L

Pane-pain merged

PANE    P EY N
PAIN    P EY N

Line-loin merged

LINE    L AY N
LOIN    L OY N

meat-meet merged

MEAT    M IY T
MEET    M IY T

happy-tensing ?

HAPPY   HH AE P IY

I think a lax happy would end with IH.

Clone this wiki locally