Topic The relationship between bipolarity and schizophrenia has been the subject of research in neurosciences, neuropsychology and cognitive psychopathology over the past few decades. This research has led to the formulation of a hypothesis suggesting that the two pathologies might be situated on a continuum. This hypothesis has been incorporated into the most recent mental disorders classifications with the introduction of the “schizoaffective disorder”, which encompasses both psychotic and thymic symptoms. In this study, we examine the hypothesis of such a continuum in the context of the pragmatic discontinuities that occur in dialogue with a psychologist and either a schizophrenic or a bipolar patient. Furthermore, the aim is to delineate the cognitive and psycholinguistic impairments observed in the schizophrenic group in comparison to the bipolar group. Methodology This study compares two matched corpora of clinical interviews with patients diagnosed with bipolar disorder and schizophrenia, conducted under identical conditions. To ensure the highest level of anonymity, the two corpora were disidentified. The interviews were conducted in a non-directive manner, with the objective of creating a conducive environment for the patients to express themselves freely, to enhance their confidence and to maintain the interaction as long as possible. During the course of the conversation, the patients were never contradicted. In order to create two balanced corpora, portions of the larger schizophrenia corpus were removed. Subsequently, an exhaustive search was conducted through inter-judge agreement to identify “discontinuities”, defined as specific breaks in a dialogue, which indicate incoherence in the dynamical structure of discourse. The concept of “discontinuity” has been previously formalized and is based on a precise definition of the patients’ involvement in the dialogue and their ability to maintain coherence in their discourse. These pragmatic discontinuities were then formally analyzed using “hierarchical modelling”, which reveals the dynamic structuring of the speech acts in a dialogic interaction and objectively highlights the discontinuities. Results A total of 25 discontinuities were identified in both corpora. The aforementioned discontinuities were observed in three bipolar conversational sequences and 11 schizophrenic ones. The sequences were then classified into categories in accordance with the principal features of their discontinuities. Firstly, three of the schizophrenic sequences were found to exhibit an “inaccessible non-literal meaning”. The sequences in question exhibit an oddity to such an extent that the meaning of the sequence is “broken” by the discontinuity. Secondly, four schizophrenic sequences were identifier and designated as “loss of the non-literal meaning”. In such instances, the schizophrenic patients exhibit a regression process whereby they return to a more basic and literal lexical content, from a rich and figurative meaning of the same lexical entry. Thirdly, two schizophrenic sequences are distinguished by the phenomenon of “semantic and/or lexical switches”. This entails a transition from a lexical entry to another, which frequently occurs multiple times. The aforementioned switches result in a loss of coherence until a discontinuity occurs. The category “semantic and/or phonological associations” comprises two schizophrenic sequences in which either a meaningless portmanteau word or an oxymoronic phrase is observed. A single bipolar sequence is associated with one of the aforementioned categories, which is also characteristic of schizophrenia. The category “semantic and/or lexical switches”. The two last bipolar sequences demonstrate, firstly, a “violation of the pragmatic use of a discursive marker” and secondly a “perturbation of the reference access”. The first sequence presents an argumentative marker “parce que” (i.e. because) which is not followed by any causes of the above. The second sequence displays a sudden shift: a highly cohesive sequence referencing a person is abruptly disrupted by a discontinuity when the bipolar patient refers to another individual. A total of 21 discontinuities were identified in the schizophrenic corpus, while the bipolar corpus exhibited four discontinuities. As a result, common pathological discourse features are proposed for consideration in relation to the two tested groups, although two distinct levels of this pragmatic disorder are clearly observable. To this end, a series of Fischer exact statistical tests were conducted on the number of relevant discourse events with respect to the number of words for each group. The statistical results demonstrate that the extension and variability of pragmatic discourse disorders are significantly more frequent and extensive in interactions with schizophrenic patients than with bipolar patients. However, we also observed that impaired psycholinguistic processes associated with the management of lexical entities and the conveyance of meaning are common to both schizophrenic and bipolar groups. Conclusion The specific hierarchical and dynamic discourse analysis methodology can be used to assist in the diagnosis of bipolar disorder and to facilitate the recognition of symptoms and prodromal signs of schizophrenia. Furthermore, it can also be employed to elucidate the psycholinguistic processes underlying the verbal expression of bipolar and schizophrenia disorders. In addition, it offers a novel perspective on the hypothesis of a continuum between schizophrenia and bipolar disorders by delineating the psycholinguistic processes that underpin discourse competence. This psycholinguistic cognitive disorder manifests the patients’ inability to regulate both the propositional content of their utterances and the semantic implications attached to them. Thus, we observe in both the two groups that the interaction is maintained “at all costs”: all patients continue the interaction, even if the cost is the discontinuity leading to incoherence. We hypothesize that maintaining the interaction leads to an increase in cognitive effort throughout the thematic unit, which can be extensive. When the cognitive effort becomes too great, the discontinuity occurs to reduce this effort and thus allow the interaction to continue until the end of the current thematic unit. This conclusion is strictly the same from the schizophrenic patients to the bipolar patients.