A battery of tests for measuring linguistic competence has been constructed. The battery consists often subtests. After validation and initial trial the tests were modified and then administered in two tryouts to schizophrenic patients and normal subjects. Test retest reliability was established in both patients and normals. In the patient group picture arrangement (story) and definitions subtests had low insignificant correlations. Among the normals household objects (categories), emotions and definitions had low correlations. On the basis of the above results definitions and emotions subtests were deleted from the test battery. On comparison of the group means patients scored lower than normals. They differed significantly only in colour naming and in MLU. This indicates that linguistic competence as measured by this instrument by and large does not deteriorate with the onset of schizophrenic illness.