This study describes the spontaneous sign language of six deaf children (6 to 16 years old) of hearing parents, who were exposed to Signed English when after the age of six they first attended a school for the deaf. Samples of their language taken at three times over a 15-month period were searched for processes and structures representative or not representative of Signed English. The nature of their developing semantics was described as the systematic acquisition of features of meaning in signs from selected lexical categories (kinship terms, negation, time expression, wh-questions, descriptive terms, and prepositions/conjunctions).
Processes not representative of Signed English were found to conform with grammatical processes of American Sign Language (ASL) and were so described. Five levels of increasing complexity of these ASL processes and of the structures representative of Signed English were hypothesized. Levels of increasing complexity of semantic feature acquisition of signs within the lexical categories named above were also hypothesized. The development of ASL grammatical processes was found to be orderly and characterized by the appearance of new processes and structures and by increases in utterance length and complexity resulting from the coordination and expansion of formerly used processes. By the end of Level 2, subjects displayed knowledge of most basic sentence forms, simultaneous processes, and some grammatical uses of repetition. A significant change appeared at Level 3, when processes that could express meaning simultaneously were used to portray communication between the signer and others as well as to express two different ideas simultaneously. These uses provided the foundation for Level 4 and Level 5 processes, which functioned to set up locations in space in increasingly complex ways. These processes decreased the need for total dependence on visual field content (i.e. immediate context), since persons and objects were first established, next assigned a position in front of the signers, bodies (i.e. set up in a particular location), and then commented about -- as opposed to the earlier process of first pointing out and then naming and describing elements of the visible "picture." Thus as the subjects became more linguistically mature, they developed "scene-setting" (typically ASL) ways of establishing who or what they were referring to, by making efficient use of signing space. This developmental change also revealed that the children were beginning to conceptualize events as wholes and to relate information about them in logical ways.
The development of structures representative of Signed English was likewise an orderly process, characterized by the appearance of new relations and the coordination and expansion of formerly used relations. By the end of Level 2, most basic semantic and grammatical relations were expressed by signs in English order. Preposition-object, appositive, genitive, and disjunctive relationst basic semantic and grammatical relations were expressed by signs in English order. Preposition-object, appositive, genitive, and disjunctive relations were later occurring relations (Levels 3'and 4), as were dative and indirect object relations (Level 5). There was some evidence to indicate that the consistent use of Signed English grammatical morphemes followed the order of acquisition of these morphemes by hearing children, presumably in both cases a function of the grammatical or semantic complexity of the morphemes themselves.
Learning the meanings for signs was, for the most part, a process in which labels (signs) for referents took on additional features and thereby became more specific. This process is compared with the general process of perception, in which initially single, usually broad or attribute features of meaning are perceived and labelled before more defining features are added to form "configurations" of features for a particular referent.
A brief contrast between the development of ASL grammatical processes and the development of structures representative of Signed English revealed both differences and similarities in development. Quite clearly, the contrast showed that the children were more linguistically competent using ASL grammatical processes -- processes be it noted for which they had no adult model.