The term “alexithymia” was introduced in the lexicon of psychiatry in the early ‘70s by Sifneos to outline the difficulties manifested by some patients in identifying and describing their own emotions. Since then, the construct has been broadened and partially modified. Today this describes a condition characterized by an altered emotional awareness which leads to difficulties in recognizing your own and others' emotions. In half a century, the volume of scientific products focusing on alexithymia has exceeded 5,000. Such an expansive knowledge domain poses a difficulty for those willing to understand how alexithymia research has developed. Scientometrics embodies a solution to this issue, employing computational, and visual analytic methods to uncover meaningful patterns within large bibliographical corpora. In this study, we used the CiteSpace software to examine a corpus of 4,930 publications on alexithymia ranging from 1980 to 2020 and their 100,251 references included in Web of Science. Document co-citation analysis was performed to highlight pivotal publications and major research areas on alexithymia, whereas journal co-citation analysis was conducted to find the related editorial venues and disciplinary communities. The analyses suggest that the construct of alexithymia experienced a gradual thematic and disciplinary shift. Although the first conceptualization of alexithymia came from psychoanalysis and psychosomatics, empirical research was pushed by the operationalization of the construct formulated at the end of the ‘80s. Specifically, the development of the Toronto Alexithymia Scale, currently the most used self-report instrument, seems to have encouraged both the entrance of new disciplines in the study of alexithymia (i.e., cognitive science and neuroscience) and an implicit redefinition of its conceptual nucleus. Overall, we discuss opportunities and limitations in the application of this bottom-up approach, which highlights trends in alexithymia research that were previously identified only through a qualitative, theory-driven approach.