The distribution of gamma-ray bursts (GRBs) detected by BATSE, in combination with the spectral data gathered over the past two decades, argue for a cosmological population of neutron star sources. We demonstrate that GRB spectra may be understood in the context of the Compton upscattering of typical radio pulsar spectra, which are steep (i.e., having a flux density Fv ∼ να with α ∼ -1 to -4) above a low-frequency cutoff νmin at ≲500 MHz. We suggest that the existence of νmin is the principal reason why GRB spectra often exhibit a break εbreak at ∼0.2-3 MeV. Due to magnetic suppression, the scattering cross section for the radio photons ∝ν2. As such, this model predicts a γ-ray power spectral index μ ∼ α + 2 + 1 above the break, though the prevalence of steeper radio spectra in the brightest pulsars implies a biasing of μ toward the bottom of this range. We find that the probability of detecting a burst in progress from any given source is ≈5 ×10-11, implying an individual stellar burst rate of about 1 every 50 yr if all active pulsars are involved, or about 1 every 2-3 yr for the very young members of this class. In addition, the energy released per burst ranges from ∼1041 ergs to ∼5 × 1043 ergs. This rate and energy release coincide with those pertaining to the macro- and microglitches seen in the periods of many such sources. We conclude that GRBs may simply be the crustal adjustments responsible for the now familiar timing noise observed in young pulsars. A simple (though important) requirement of this model - that εbreak should be correlated with the burst luminosity - seems to be in agreement with the recent analysis of SIGNE data.