Recently, considerable attention has been paid to Bombyx mori silk fibroin by a range of scientists from polymer chemists to biomaterial researchers because it has excellent physical properties, such as strength, toughness, and biocompatibility. These appealing physical properties originate from the silk fibroin structure, and therefore, structural determinations of silk fibroin before (silk I) and after (silk II) spinning are a key to make wider applications of silk. There are discrepancies about the silk I structural model, i.e., one is type II β-turn structure determined using many solid-state and solution NMR spectroscopies together with selectively stable isotope-labeled model peptides, but another is α-helix or partially α-helix structure speculated using IR and Raman methods. In this review, firstly, the process that led to type II β-turn structure by the authors was introduced in detail. Then the problems in speculating silk I structure by IR and Raman methods were pointed out together with the problem in the assignment of the amide I band in the spectra. It has been emphasized that the conformational analyses of proteins and peptides from IR and Raman studies are not straightforward and should be very careful when the proteins contain β-turn structure using many experimental data by Vass et al., (Chem. Rev., 2003, 103, 1917–1954). In conclusion, the author emphasized here that silk I structure should be type II β-turn, not α-helix.