Most educators know something about the language development of native English-speaking children. There is a vast literature on the topic, and the benchmarks of language development have been plotted for monolingual children; however, no such benchmarks are available for bilingual children. Children follow different paths to become bilingual and the stages that they pass through can be quite varied.

Stages of Development

Currently researchers believe that there is a consistent developmental sequence that children follow in acquiring a first language. If the child acquires two languages simultaneously, the stages of development are the same as they are for monolingual speakers of those languages. There is debate over whether bilingualism results in a slower rate of vocabulary development than is true of children learning the same languages monolingually. Goodz (1994) reports no delay or retardation, but other researchers have reported lower vocabulary scores for bilingual than for monolingual children in a given language (Bialystok, 1988; Doyle, Champagne, and Segalowitz, 1978).

Typically, children who are learning two languages simultaneously make unequal progress in the languages. One language is more salient from time to time, either because of the input that the child is receiving from other speakers, or because there are more opportunities to use one language than the other. However, there is no simple relationship between a child's proficiency in each language and the amount of input in that language from caregivers and others (Goodz, 1994).
For children who are learning a second language sequentially, the development progression is somewhat different. Tabors and Snow (1994) argue that such children pass through four distinct stages:

1. First, the child uses the home language. When everyone around the child is speaking a different language, there are only two options-to speak the language they already know, or to stop speaking entirely. Many children, but not all, follow the first option for some period of time (Saville-Troike, 1987). This of course leads to increasing frustration, and eventually children give up trying to make others understand their language.

2. The second stage is the nonverbal period. After children abandon the attempt to communicate in their first language, they enter a period in which they do not talk at all. This can last for some time, or it can be a brief phase. Although they do not talk during this time, children attempt to communicate nonverbally to get help from adults or to obtain objects. Furthermore, this is a period during which children begin actively to crack the code of the second language. Saville-Troike (1987) noted that children will rehearse the target language by repeating what other speakers say in a low voice and by playing with the sounds of the new language.

3. The next stage occurs when the child is ready to go public with the new language. There are two characteristics to this speech-it is telegraphic and it involves the use of formulas. Telegraphic speech is common in early monolingual language development and involves the use of a few content words without function words or morphological markers. For example, a young child learning to speak English may say "put paper" to convey the meaning, "I want to put the paper on the table." Formulaic speech refers to the use of unanalyzed chunks of words or routine phrases that are repetitions of what the child hears. Children use such prefabricated chunks long before they have any understanding of what they mean (Wong Fillmore, 1976).

4. Eventually, the child reaches the stage of productive language use. At this point the child is able to go beyond short telegraphic utterances and memorized chunks. Initially, children may form new utterances by using formulaic patterns such as "I wanna" with names for objects. In time, the child begins to demonstrate an understanding of the syntactic system of the language. Children gradually unpackage their formulas and apply newly acquired syntactic rules to develop productive control over the language.

Like any scheme of developmental stages, the sequence outlined here is flexible. At a given stage children have recourse to previously used strategies. Formulaic speech is still used in the stage of productive language use, for example. Rather than speaking of stages, it makes more sense to speak of waves, in that waves can be visualized as moving in and out, generally moving in one direction, but receding, then moving forward again (Olsen Edwards, personal communication, 1994). This seems to capture more accurately the child's development-in language and in other areas as well.

Furthermore, there are vast individual differences with respect to the rate at which children pass through the different stages. Some children go through a prolonged nonverbal stage, sometimes lasting for a year or more, whereas other children pass through this stage so quickly they seem to have rejected this strategy altogether. Nora, in Wong Fillmore's (1976) study, preferred to interact with English-speaking children and used every opportunity to engage in meaningful conversation in that language. Other learners in the same study chose to speak almost entirely with other children who understood their first language and so made little progress in the second language.

Language Mixing

Most observers of children learning two languages simultaneously note that there is some mixing of languages at the lexical level. There is a great deal of controversy about how much mixing occurs and what it means. Recent research by Goodz (1994) suggests that mixing increases somewhat during early childhood, peaking at 30 months or so, and then declining. She followed 13 children and their parents and focused on the input the parents provided. In spite of parents' protestations to the contrary, observations indicated that they did not separate languages by person; rather, in all cases there were situations when parents used their nonnative language with the child. This research was done with French-English speakers in Canada; research with Mexican-American families in the United States indicates mixing is quite common in some communities.

In such communities mixing languages and switching from one language to another is part of the child's normal linguistic environment. Language mixing and code switching are used for definite communicative needs. Speakers build on the coexistence of alternate forms in their language repertory to create meanings that may be highly idiosyncratic and understood only by members of the same bilingual speech community. In such communities adult code-switching is a rhetorical strategy used in such communicative tasks as persuading, explaining, requesting, and controlling. It is preferred to other rhetorical devices because it has greater semantic power deriving from metaphorical allusion to shared values and to the bilinguals' common problems vis-a-vis the society at large.

A number of observers have noted that when bilinguals have been interacting mainly with other bilinguals for a long time, the model for each of their languages is not monolingual usage of these languages but rather the languages as spoken by the bilinguals themselves. In these situations, the mixed speech becomes a code of its own-"contact language" (Haugen, 1953) that is used to stress in-group behavior or emphasize informality or rapport.

It is important for educators in early childhood education programs to realize that language mixing (inserting single lexical items from one language into another) and code switching (switching languages for at least a phrase or a sentence)-are common linguistic devices in many of the communities from which their students come. Rather than indicating that children are confusing their two languages, such phenomena can be a sign of linguistic vitality. Young children in such communities are in the process of learning to switch languages in the sophisticated manner they hear around them. Teachers who switch languages are merely adjusting their speech to the language of the child's community and culture.

Language Loss and Semilingualism

It sometimes happens that children lose their first language skills as the second language begins to predominate (Wong Fillmore, 1991). Because of the emphasis put on English in our schools and society, children can gradually lose aspects of their first language. This is sometimes the case in immigrant families, where the parents are learning English and that language begins to predominate in the home.
Because, developmentally, children may be losing aspects of their first language as the second language is acquired, their performance on tests of language proficiency can be misleading. At a given point in time, their skills in both languages may be relatively weak. This has led some researchers to talk of "semilingualism," a condition where children are not functioning well in either language. Indeed, one often hears teachers decrying the fact that bilingual children "don't know either language."
Most linguists would restrict the use of the term "semilingualism" to those cases where extreme social deprivation results in bilingual children not functioning well in either language. Although instances of extreme linguistic and communicative deprivation may lead to language pathology, usually what appears to be semilingualism is only a temporary phase in language development. Thus, as shown in Figure 1, there may be a developmental period when lack of use of the first language results in a decline in proficiency while at the same time knowledge of the second language is not yet at an age-appropriate level.

FIGURE 1
Semilingualism as a function of declining proficiency in the first language and less than adequate proficiency in the second
(Note: Due to the constraints of the electronic environment Figure 1 has been omitted.)

The concept of semilingualism is not a useful way to refer to this developmental phase because, even though a bilingual child's performance in either language may lag behind that of monolingual speakers of the language at some point in development, the child may actually possess a total vocabulary and total linguistic repertory that is quite similar to that of monolingual speakers. Rather, this occurrence may be more appropriately described as a language imbalance, where at certain points in the development of their languages bilingual children do not perform as well as native speakers in either language. There may be shifts back and forth in which language is more proficient, depending on language use and exposure, but most bilingual children are able eventually to come up to age-level proficiency in both languages given more exposure and opportunities for use of the weaker language.

To summarize, bilingual language development can follow a number of different patterns. Especially for sequential bilinguals (Types 3 and 4), there is often imbalance in the child's languages as one or the other language predominates. For certain children language mixing and code-switching are part of the linguistic repertory. Educators in early childhood education programs need to be sensitive to these complexities in bilingual language development. This does not mean that language assessment is impossible, but it does mean that extra sensitivity is required.