Automatic speech recognition is expected to be more successful when syntactically-related information is incorporated into early stages of recognition. Phonemic decisions, in particular, are expected to be more accurate and less ambiguous when contextual information is considered. A computer program detected about 90% of all boundaries between major syntactic constituents from fall-rise fundamental frequency (Fo) contours in the Rainbow Script as read by two talkers. A procedure was devised for locating stressed syllables, in constituents, from (1) high-energy and increasing-Fo portions near the peaks of Fo contours and (2) local increases in Fo from an archetype falling contour. The procedure succeeded in locating most syllables which had been perceived by listeners to be stressed. A recognition strategy is outlined for using detected syntactic boundaries and stressed syllable locations in estimating distinctive features of some phonemes in connected speech. Vowel and consonanat recognition would be attempted first in the stressed syllables. Other readily-detected segments, such as coronal strident fricatives, would be found. A pilot study showed that front/back decisions for vowels are more reliable in stressed than unstressed syllables. (Author/DD)