Course materials for/by Peter L. Patrick. May contain copyright material used for educational purposes. Please respect copyright.


An Introduction to

the (TD) Sociolinguistic Variable

Peter L Patrick, Univ. of Essex


These notes serve to briefly introduce the phonological linguistic variable of final /-t, -d/ deletion (consonant cluster simplification) in varieties of English, describing several of the major linguistic constraints that have been found to systematically govern variation across all varieties of English.

A bibliography with comments is available here. A table comparing frequencies of 13 classic studies for the two major linguistic constraints is available here (from Patrick 1999: 162).


1. Definition of the Variable:

(TD)-deletion in English is the deletion of apical (=alveolar) stops /t/ and /d/ in final clusters. You can hear (TD)-Deletion at work in such examples as:

        “I don’ know”   

        “Then he pass’ me his plate”   

        “She tol’ a lie”

It’s a naturally variable reduction process that all speakers of English do, usually unconsciously, in frequencies that differ according to a variety of factors:

Linguistic, e.g. the surrounding sounds or type of word it occurs in;

and  Social, such as speaking style, and speaker sex, age, class, ethnicity.


(TD)-deletion studies look at the presence or absence of a final /-t/ or /-d/ element in consonant clusters where it is underlying, and thus expected to occur. Since it sometimes occurs and sometimes does not, the principal analytical question is:

        “What factors favor the deletion of final /-t/ or /-d/?”

The answers are usually quite predictable in a general way, but allow for subtle and important differences of social, dialectal and individual patterning. Thus, a detailed, quantitative profile of these factors can tell us something about how closely groups of speakers resemble each other.



2. Exclusions from the Data:

Due to the nature of the variable there are cases which cannot usefully be looked at in a variationist, quantitative analysis. These are called the “Don’t-Count” cases, e.g.

·        Cases where the /t,d/ that might be deleted is followed by another /t/ or /d/:

You can’T Tell                  /yuw kaent tEl/

·        Cases where preceding /r/ is a crucial member of the consonant cluster:

Which paRT is it               /wItsh part Iz It/

·        Cases where a nasal before the (TD) and a vowel after it allow the rule of “nasal-flapping”:

Wanna (WaNT A) light?     /wa?a lait/


In collecting our corpus of tokens we systematically excluded these cases.



3. Explanatory or Independent Linguistic Variables:

(these are often called “Factor Groups” in VarbRul analysis)


The first type of linguistic factor to consider is phonological environments. Preceding Segment refers to the class of sounds that occur before the (TD), that is, those sounds that make up the first part of the cluster. (Naturally, they’re limited to consonants!)


1) Sibilants     /s, sh, ch, z, zh, j/   Ex.:  paST, waSHED, raZZED, parCHED

2) Stops           /k, p, g, b/               Ex.:  raPT, baGGED, baKED, naBBED

3) Fricatives    /f, v, th, Th/             Ex.:  raFT, depraVED, baTHED

4) Nasals         /n, m, ng/                Ex.:  waNT, waNED, baNGED, taMED

5) Liquids        /l/                            Ex.:  oLD, haLT, caLLED



Following Segment refers to the class of sounds that follow the (TD) -- the sounds that make up the first part of the following morpheme (including pause, if one occurs).

1) Consonant (all groups above)    Ex.:  wanT Some, bakeD Goods

2) Glides         /y, w, h/                   Ex.:  stanD Here, calleD You

3) Rhotics       /r/                           Ex.:  tolD Rita, planneD Reunion

4) Vowel          (all Eng. vowels)       Ex.:  calleD A cab, warneD Us

5) Pause          (i.e., silence)            Ex.:  the enD…, are you colD?



The second type of linguistic factor to consider is Grammatical Category. This refers to the type of morpheme the (TD) cluster occurs in:


1) N’T      (negative-suffix items)    Ex.:  caN’T, doN’T, isN’T, aiN’T

(N’T is the most-often-deleted type; some studies exclude it, because of this predictable variation in frequency.)

2) Send/Sent        (Irreg. devoicing verbs)    Sent/Spent/Lent/Bent/Built

(“Send”-type verbs are quite rare, so they are often not considered either.)

3) Mono-morphemes     (/-t, -d/ part of stem)      Ex.: olD, bursT, denT

4) Semi-weak         (Verbs w/vowel-change+TD)   Ex.: tol-D,kep-T,mean-T

5) Past     (Regular past verbs)         Ex.: pass-ED, bill-ED, tone-D


Bibliography of the (TD) variable

More links about Language Variation

Back to Peter L. Patrick's homepage

Last updated on 17 November 2003