State-of-the-art computational methods in solid-state chemistry were applied to predict the structural and spectroscopic properties of the α and γ crystalline polymorphs of nylon 6. Density functional theory calculations augmented with an empirical dispersion correction (DFT-D) were used for the optimization of the two different crystal structures and of the isolated chains, characterized by a different regular conformation and described as one-dimensional infinite chains. The structural parameters of both crystalline polymorphs were correctly predicted, and new insight into the interplay of conformational effects, hydrogen bonding, and van der Waals interactions in affecting the properties of the crystal structures of polyamides was obtained. The calculated infrared spectra were compared to experimental data; based on computed vibrational eigenvectors, assignment of the infrared absorptions of the two nylon 6 polymorphs was carried out and critically analyzed in light of previous investigations. On the basis of a comparison of the computed and experimental IR spectra, a set of marker bands was identified and proposed as a tool for detecting and quantifying the presence of a given polymorph in a real sample: several marker bands employed in the past were confirmed, whereas some of the previous assignments are criticized. In addition, some new marker bands are proposed. The results obtained demonstrate that accurate computational techniques are now affordable for polymers characterization, opening the way to several applications of ab initio modeling to the study of many families of polymeric materials.