The contributions of vocal‐fold tension and of subglottal air pressure to changing fundamental frequencyF0are indirectly assessed by comparing human glottal‐source data with synthetic glottal waveforms generated by the Ishizaka–Flanagan two‐mass model of vocal‐fold vibration. In this model, synthetic male and female glottal waves may be generated with known values of subglottal air pressure and vocal‐fold tension. Human glottal waves were recorded with a reflectionless tube into which ten male and female adults phonated. The subjects produced interrogative (risingF0) and declarative (fallingF0) glides and trisyllabic words with primary stress on the initial, medial, or final syllable. Two types of change in the glottal wave over time are possible, depending on whetherF0is changed primarily by vocal‐fold tension or by subglottal air pressure. Comparison of the natural and synthetic glottal waves indicates that (1) the rise of frequency in interrogative words is due principally to increasing vocal‐fold tension, while (2) the fall of frequency in declarative words is due principally to decreasing subglottal air pressure; (3) in the polysyllabic words, the change of frequency within syllables resembles that of the declarative monosyllables and appears due primarily to changes of subglottal air pressure; and (4) the heightenedf0of the stressed syllable is due to an increase in the vocal‐fold tension, typically accompanied by increased subglottal air pressure.