With the advent of von Neumann-style computers, widespread exploration of new methods of music composition became possible. For the first time, complex sequences of carefully specified symbolic operations could be performed in a rapid fashion. Composers could develop algorithms embodying the compositional rules they were interested in and then use a computer to carry out these algorithms. In this way, composers could soon tell whether the results of their rules held artistic merit. This approach to algorithmic composition, based on the wedding between von Neumann computing machinery and rule-based software systems, has been prevalent for the past thirty years. The arrival of a new paradigm for computing has made a different approach to algorithmic composition possible. This new computing paradigm is called parallel distributed processing (PDP), also known as connectionism. Computation is performed by a collection of several simple processing units connected in a network and acting in cooperation (Rumelhart and McClelland 1986). This is in stark contrast to the single powerful central processor used in the von Neumann architecture. One of the major features of the PDP approach is that it replaces strict rule-following behavior with regularity-learning and generalization (Dolson 1989). This fundamental shift allows the development of new algorithmic composition methods that rely on learning the structure of existing musical examples and generalizing from these learned structures to compose new pieces. These methods contrast greatly with the majority of older schemes that simply follow a previously assembled set of compositional rules, resulting in brittle systems typically unable to appropriately handle unexpected musical situations. To be sure, other algorithmic composition methods in the past have been based on abstracting certain features from musical examples and using these to create new compositions. Techniques such as Markov modeling with transition probability analysis (Jones 1981), Mathews' melody interpolation method (Mathews and Rosler 1968), and Cope's EMI system (Cope 1987) can all be placed in this category. However, the PDP computational paradigm provides a single powerful unifying approach within which to formulate a variety of algorithmic composition methods of this type. These new learning methods combine many of the features of the techniques listed above and add a variety of new capabilities. Perhaps most importantly, though, they yield different and interesting musical results. This paper presents a particular type of PDP network for music composition applications. Various issues are discussed in designing the network, choosing the music representation used, training the network, and using it for composition. Comparisons are made to previous methods of algorithmic composition, and examples of the network's output are presented. This paper is intended to provide an indication of the power and range of PDP methods for algorithmic composition and to encourage others to begin exploring this new approach. Hence, rather than merely presenting a reduced compositional technique, alternative approaches and tangential ideas are included throughout as points of departure for further efforts.