A dialect refers to any variety of language that is shared by a group of speakers. It is not possible to speak a language without also speaking a dialect (Wolfram and Schilling-Estes, 1998). Although all dialects of a language are equally systematic and complex, on a social level, dialects are often described as falling on a continuum of standardness. The most standard dialect of a language generally reflects an idealized prestige form that is rarely spoken by anyone in practice. Rules for producing this standard, however, can be found in formal grammar guides and dictionaries. Versions of the standard can also be found in formal texts that have been written by established writers. Next in standardness are a number of formal and informal oral dialects. These dialects reflect the language patterns of actual speakers. Norms of acceptability for these dialects vary as a function of the regional and social characteristics of different communities and of different speakers within these communities. Nonstandard dialects represent the other end of the continuum. These dialects also reflect spoken language, but they include socially stigmatized linguistic structures. Other terms used to describe nonstandard dialects are nonmainstream and vernacular.

At the linguistic level, scholars repeatedly highlight the arbitrary nature of a dialect's social acceptability (i.e., standardness). In fact, Milroy and Milroy (2000) argue that contradictory and changing attitudes to the same linguistic phenomenon can emerge at different times in the history of a language. For example, as these authors note, before World War II, absence of postvocalic /r/ in words such as car and park was not stigmatized in New York City. By 1966, however, r-lessness had become a stigmatized marker of casual style and lower social class. English dialects containing r-lessness continue to be stigmatized in the United States, but in England, English dialects with this same linguistic pattern have high status.

Most linguistic patterns that occur in the standard dialects of a language also occur in those that are nonstandard. Seymour, Bland-Stewart, and Green (1998) refer to these language patterns as noncontrastive, because all dialects of a language are thought to share these forms. Despite the similarity that exists among dialects, nonstandard versions are typically described by listing only those language patterns that do not appear in the standard varieties. Seymour, Bland-Stewart, and Green refer to these patterns as the contrastive features. Descriptions of nonstandard dialects become even narrower when they are generated by the media and general public, because these descriptions tend to highlight only the contrastive patterns that are highly stigmatized (Rickford and Rickford, 2000). Zero marking of the copula be (i.e., he walking) is an example of a pattern that is frequently showcased for African-American English (AAE) (Rickford, 1999). Other stereotypic patterns include the use of ya'll and fixing to to describe versions of Southern White English (SWE), a-prefixing (e.g., he was a-walking …) to characterize Appalachian English, and pronouncing think and that as tink and dat to depict Cajun English.

Although it is relatively easy to identify the language forms that differentiate a nonstandard dialect from one that is viewed as standard, it is much more difficult to identify patterns that distinguish one nonstandard dialect of a language from another. One reason for this is that many nonstandard dialects of a language share the same contrastive patterns. Unique contrastive patterns for different nonstandard dialects are particularly rare when the dialects being compared are produced in the same community and by speakers of the same social class. Oetting and McDonald (2001, 2002) illustrated this finding by comparing the contrastive patterns of two nonstandard dialects spoken in southeastern Louisiana. The data for this comparison were language samples of children who lived in the same rural community and attended the same schools. Forty of the children were African American and spoke a southern rural version of AAE, and 53 were white and spoke a rural version of SWE. The AAE and SWE dialects spoken by the children were deemed distinct through the use of a listener judgment task and a discriminant function analysis of 35 different nonstandard (i.e., contrastive) language patterns found in the transcripts. Nevertheless, of the 35 contrastive patterns examined, 31 of them were found in the conversational speech of both the AAE and SWE child speakers.

Besides contrastive forms, there are three other ways in which dialects differ from one another. One way is in the frequency with which particular language forms are produced. Wolfram's (1986) analysis of consonant cluster reduction in 11 different English dialects is useful for illustrating this finding. His data showed cluster reduction occurring 3% of the time for standard English and northern white working-class English, 4% for northern African-American working-class English, 10% for southern white working-class English, 36% for southern African-American working-class English, 5% for Appalachian working-class English, 10% for Italian-American working-class English, 22%–23% for Chicano working-class and Puerto Rican working-class (New York City) English, 60% for Vietnamese English, and 81% for Native American Puebloan English. For each dialect listed, the percentage reflects the degree to which consonant clusters were reduced in regular past tense contexts that were followed by a nonconsonant (e.g., “Tom live in”). As can be seen, all 11 of the dialects showed cluster reduction, but the frequency with which this pattern occurred in each dialect greatly varied.

A second way in which dialects differ from one another has to do with the linguistic environments in which particular language forms occur. The effects of linguistic context on language use are often described as linguistic constraints or linguistic conditions (Chambers, 1995). Studies of linguistic constraints typically occur with language forms that show systematic variability in their surface structure. An example of a variable form is one that can be overtly expressed (i.e., present in the surface grammar; he is walking) in some contexts but is zero-marked (i.e., absent from the surface grammar; he walking) in others.

One of the most widely studied variable forms is the copula be, and research on this structure has typically involved nonstandard versions of AAE. At least six different linguistic constraints have been found to influence AAE speakers' use of the copula be. These constraints include the type of preceding noun phrase (noun phrase versus personal pronoun versus other pronoun), the phonological characteristics of the preceding environment (vowel versus consonant; voiced versus unvoiced consonant), the person, number, and tense of the verb context (first, second, third; present versus past), the grammatical function of the be form (copula versus auxiliary), the nature of the following predicate clause (locative versus adjective versus noun), and the phonological characteristics of the following environment (vowel versus consonant) (Rickford, 1999). The person, number, and tense of the verb context, the grammatical function of the be form, and the nature of the following predicate clause also have been shown to affect copula be marking in various nonstandard SWE dialects (Wolfram, 1974; Wynn, Eyles, and Oetting, 2000). Other morphosyntactic structures of English that have been found to be influenced by linguistic constraints include, but are not limited to, negation, do support, verb agreement, relative pronouns, plural marking, and question inversion (Poplack, 2000).

A third way in which dialects differ from one another is in the semantic meanings or grammatical entailments of some forms (Labov, 1998). These particular cases involve language patterns that occur in most dialects of a language, but their meanings or use in the grammar are unique to a particular dialect. These patterns are often described as camouflaged forms because, on the surface, the contrastive nature of these forms can be difficult to notice (Wolfram and Schilling-Estes, 1998). Use of had + Ved is an example of a camouflaged pattern. In most dialects of English, this structure carries past perfect meaning (e.g., “I already had eaten the ice cream when she offered the pie”). In some English dialects such as AAE, however, had + Ved also can express preterite (i.e., simple past) meaning (Rickford and Rafal, 1996). Ross, Oetting, and Stapleton (in press) provide the following sample from a 6-year-old AAE speaker to illustrate the preterite meaning of this form: “Then my mama said, `It's your mama. Let me talk to your daddy.' Then she had told my daddy to come with us and bring a big rope so they could pull the car home.”

Dialect use is affected by factors that are both internal and external to a speaker (Milroy, 1987; Chambers, 1995; Wolfram and Schilling-Estes, 1998). Internal factors that have been shown to influence the type and density of one's dialect include age, sex, race, region of the country, socioeconomic status, type of community, and type of social network. Interestingly, regardless of race, region, community, and network, members of lower social classes produce a greater frequency of contrastive dialect forms than members of higher classes. Greater frequencies of contrastive patterns also have been found for younger adults than for older adults, and for males than for females. Exceptions to these generalities do exist, however. For example, Dubois and Horvath (1998) documented a V-shaped age pattern rather than a linear one in their study of Cajun English. They also found that the type and degree of the age pattern (V-shaped versus linear) depended on the speaker's sex and type of social network. In particular, the V-shaped pattern was more pronounced for men than for women, and only women from closed social networks showed the V-shaped pattern. Women in open networks showed a linear pattern. Interestingly, though, the linear trend reflected higher frequencies of nonstandard dialect use by the older women than younger women. This finding contrasts with what is typically reported in other nonstandard dialect work, namely, that older adults present fewer instances of nonstandard forms than younger adults.

Some of the external factors that have been shown to affect dialect use include the type of speaking style (casual versus formal; interview versus conversation), speaking partner (familiar versus unfamiliar; with authority versus without), modality of expression (speaking versus writing versus reading), genre (persuasive versus informative versus imaginative), and type of speech act (comment versus request for information) (for data, see Farr-Whitman, 1981; Labov, 1982; Smitherman, 1992; Lucas and Borders, 1994). Influences of these external factors on dialect use interact in complex ways with those that are internal to a speaker. The dynamic interactions that occur between and among these variables help explain why dialect use is often described as fluid, flexible, and constantly changing.

The challenge for scientists in communication disorders is to learn how a speech or language disorder affects a person's use of language, regardless of the dialect spoken. Thus far, most descriptions of childhood language impairment have been made within the context of standard dialect varieties only. Extending the study of childhood language impairment to different nonstandard dialects is a topic of current scholarly work and debate (see dialect versus disorder).