This article explores the intensity of processing during sentence comprehension by measuring pupillary response during reading. Two experiments contrast the processing of simpler versus more complex sentences. The two more complex sentence types, object-relative center-embedded sentences and filler-gap sentences, not only take longer to process than their simpler counterparts, but they also produce a larger change in pupil diameter. We propose that the pupillary response is an indicator of how intensely the processing system is operating. The more complex sentences evoke some intense processing at the point in the sentence where a syntactic complexity is first encountered. The gaze durations at these points are elevated, indicating the immediate response to the demand of the syntactic processing. The pupil then starts to dilate, reaching a maximal diameter approximately 1.3 s later. The results from these various performance measures are integrated within a resource-limited computational model of comprehension. The paper develops a model of comprehension that includes an intensity dimension of thought, drawing a correspondence between the computational model's consumption of resources and the human pupillary response.