Optical signals generated in the interaction zone between an Nd:YAG laser beam and an aluminum alloy have been investigated, in order to gain an appreciation of their potential for use in adaptive control systems for laser processing. A system has been developed that enables such signals to be recorded and analyzed. The laser beam welding process has been monitored, from the first visible signs of surface damage through to full penetration of the work piece, by linearly increasing the laser power during a melt run performed with a constant traverse rate. Trials were carried out on samples of the aluminum alloys AA5083 and AA5005 with various surface preparations and with a variety of sensor configurations. The characteristics of the resultant welds were established by using optical and scanning electron microscopy on transverse weld sections. The signals measured were consistent with well-defined characteristic modes of interaction. The effects of surface penetration, surface reflectance and weld pool curvature on the signals produced are noted and discussed. A fundamental understanding of the potential for optical signals to be used for on-line characterization of laser-material interactions was achieved through analysis and validation of the data captured. The methods used are sufficiently generic to be applied to other materials and laser-based processes.