We resolve information about the dynamics of simple liquids that has been obscured in prior frequency and time-domain measurements by way of full field-resolved polarization emission time (FR-PET) measurements of carbon disulfide. The amplitude and phase of the field-resolved transient-grating signal is used to calculate a spectrogram of the signal field at each delay between the transient-grating (TG) pump and probe pulses. The temporal maximum of the spectrogram, defined to be the signal emission time, varies with pump-probe delay; it follows the convolution of the TG pulses while the pulses overlap and exhibits recurrences at times when the nuclear dynamics are the main component of the liquid material response. Since this is a third-order nonlinear spectroscopic method, the isotropic and anisotropic signals are constructed from the polarization tensor components. The frequency-integrated anisotropic component of the signal is equivalent to the signal measured in optical Kerr Effect (OKE) experiments. The FR-PET determination of the signal emission times is a direct measurement of the third-order nonlinear (polarizability) polarization and, hence provides new strong constraints on appropriate models of the liquid dynamics. Models for the material response function are used to calculate the signal emission times. In particular, we show that the proper treatment of the time-correlation function for orientational motion gives the best fit to the FR-PET data for rotational diffusional motion. We also establish that librational motion is not a short-time (coherent) motion that leads to rotational diffusion. Finally, we find that the Bucaro-Litovitz form for interaction-induced dynamics is not entirely correct for the CS(2) liquid we study. We suggest that the failing may result from the implicit assumption of two-body interactions, which is only appropriate for gases.