A debate has simmered concerning the nature of clinical reasoning, especially diagnostic reasoning: is it a "science" or an "art"? The trend since the seventeenth century has been to regard medical reasoning as scientific reasoning, and the most advanced clinical reasoning is the most scientific. However, in recent years, several scholars have argued that clinical reasoning is clearly not "science" reasoning, but is in fact a species of narratival or hermeneutical reasoning. The study reviews this dispute, and argues that in a theoretical sense, the dispute rests upon a naïve--but very popular--caricature of what constitutes "science reasoning." But, if the dispute rests upon just such a caricature, why is it so persistent? The study concludes by suggesting that we, as patients and as physicians, have deep psychological tendencies that incline us to adopt the very naïve "science" concept/model of diagnostic reasoning, even if (or when) we understand its inaptness.