The human language system comprises a series of processing stages across multiple brain regions typically lateralized to the left cerebral hemisphere. In the event of stroke or other injury to these systems, neural circuits reorganize in an attempt to restore lost function. However, the neural processes and learned representations underlying language network reorganization remain unclear. Here, we present data from a first-of-its-kind human case study involving chronic intracortical recordings from the right cerebral hemisphere of a patient with non-fluent aphasia with severe left-hemispheric damage to frontal and parietal cortices. To explore the extent to which right-hemispheric language homolog areas contribute to recovery and compensate for stroke-induced loss of function in left-hemispheric language networks, we analyzed large-scale extracellular recordings from multi-electrode arrays implanted in regions of the language network known to be important for language comprehension and language production. Neuronal activity was recorded while the patient performed linguistic tasks designed to probe language comprehension, language production (picture naming), and language repetition. In line with non-fluent aphasia, the patient exhibited in particular difficulties with word-retrieval during naming but displayed good performance in comprehension and repetition. Recording results from multiple experimental sessions indicated right-hemisphere engagement in language functions that are both brain region and task specific. Mean individual neuron activity across recorded areas showed distinct dynamics spanning the preparation and speech production epochs. Notably, these effects appeared consistent across tasks, suggestive of a production-intrinsic effect. Despite similarities at stimulus onset, correct and incorrect trials yielded distinguishable area-specific responses during both subsequent epochs. This distinction across areas suggests a unique, differentiable involvement of the recorded right-hemispheric brain regions in speech planning and word retrieval. In summary, our results expose a remarkably rich and interesting dataset exploring language function at both a single-neuron and neural population level while pointing to distinguishable roles of different brain regions in controlling rehabilitated speech in an adult human.