Our work on computerized analysis of the 2-channel, 24-hr electrocardiogram has resulted in the development of multi-channel signal processing systems that learn by observation. In this paper a new tool for implementing such algorithms is described: the pattern recognition language $\cal SEEK$. Programs written in $\cal SEEK$ build a knowledge base containing tree-like data structures, each of which stores acquired information about a particular multi-channel waveform. Input data is interpreted by performing an efficient parallel evaluation of the structures in the knowledge base. Our work is applicable to a wide variety of pattern recognition problems that arise in medical signal processing. The approach is illustrated with examples drawn from ECG analysis.