Introduction. The focus of the present paper is on (1) how dyslexia research and hence definitions have developed during the period 1950–2020 and includes (2) a database search of scientific publications on dyslexia during the same period. The focus is on the definitions of dyslexia and the organization of the network search based on the causal four-level model by Morton and Frith. Method. (1) The definitions are presented in accordance with a historic review of dyslexia research from 1950 to 2020 and based on (2) Google Scholar counts of publications on dyslexia, on defining dyslexia, on dyslexia at the four levels (symptomatic, cognitive, biological, environmental), and by areas (sensorimotor, comorbidity). Finally, a percentage calculation shows the relative development within each level and area by decennium (1950–1960, 1960–1970, 1970–1980, 1990–2000, 2002–2010, 2010–2020). Results. (1) Of the seven definitions presented, only the definition by the BDA 2007 included the four levels of the causal model. (2) The number of publications increased substantially over the period. However, relatively few publications have defined dyslexia. An increase in publications from 1950 to 2020 was seen across the four levels and two areas—however, with an alteration in the thematic focus over this time span. Summary. Defining dyslexia has still not reached a consensus. This uncertainty may explain why only one of the seven definitions proved satisfactory according to the four-level model. Along with the general increase in research, publications on dyslexia have increased accordingly during the period 1950 to 2020. Although the symptomatic level has played a dominant role over the whole period, thematic shifts have been seen over these 70 years. In particular, a substantial thematic shift was seen by the turn of the millennium. There has been a relative increase in the focus on literacy at the symptomatic level, on phonological awareness at the cognitive level, in gender at the biological level, and second language learning as comorbidities. However, increases in counts are not alone a valid indication of scientific progress. In particular, the lack of definitional criteria as a basis for participant and method selection should attract much more focus in future studies. The present study underlines the multifactorial nature of dyslexia, as evidenced by a substantial increase in the number of publications on the subject. It is a challenge for future research to continuously use and possibly redefine dyslexia definitions in line with such standards.