Research into the rate of whole-body oxygen consumption (VO2) kinetics during exercise increments to low- to moderate-intensity steady-state exercise was originally based on the theory of linear first-order VO2 kinetics, implying that the VO2 response to steady-state exercise increments is a mono-exponential response of the same time constant (tau, τ) across all steady-state intensities. Despite the acceptance of this theory for more than 30 years, early research from the 1980s documented an increasing τ with increasing steady-state exercise intensity, and recent research has confirmed such results. Today, such evidence has led to retraction of the theory of linear first-order VO2 kinetics. This history, revealing the premature acceptance of a theory, and subsequent scientific investigation using improved research design, instrumentation and data processing, has important implications for the fragility of scientific theories and the need for continual testing of theories in the search for facts and not prematurely accepted constructs. This review provides historical evidence for a critical reappraisal of the theory of linear first-order VO2 kinetics and presents data to show the need for changes in the data-processing ‘standards’ of the discipline to improve measurement of instantaneous VO2 kinetics and the time to steady state. For example, to date, no study of VO2 kinetics has quantified and statistically analysed the time to steady state. Furthermore, the instability of τ across different exercise increments, and for the same increment from different baseline VO2 demand, prevents τ from being a valid measure of VO2 kinetics for different exercise conditions. The concept of quantifying kinetics from a total non-linear response, when no other field of kinetics pursues this methodology, also raises concern for the methods and models used to interpret steady-state VO2 kinetics.