Cell to cell signaling is fundamentally combinatorial, as many signals are integrated to control and regulate the cellular response. The bone morphogenetic protein (BMP) signaling pathway, is a key example, as it comprises over 15 ligands that interact promiscuously with multiple receptors, but all phosphorylating the same downstream targets. Together, they regulate many biological processes throughout embryonic development, as well as during homeostasis. These ligands and receptors generally appear in various combinations, across most contexts, with receptors being expressed by many cell types and many environments contain multiple activating ligands. Naively, this would result in a strong, nonspecific response across cell types, in contrast to the specific temporal and spatial control needed for proper development. To resolve this conundrum, we combine quantitative experimental approaches together with mathematical modeling in order to understand how ligands combine together to activate the BMP pathway. We find that the BMP pathway provides a computational network capable of interpreting ligand combinations. We further use this insight in order to identify a new mechanism in which promiscuous signaling networks can utilize combinations of ligands to activate specific target cell types. Together, these results provide a new framework to control and regulate cellular processes. Such quantitative understanding of the way cells extract information from the signaling environment will enable new therapeutic approaches, as well as new strategies for tissue engineering.