In this article, we propose a novel form of unsupervised learning, continual\ncompetitive memory (CCM), as well as a computational framework to unify related\nneural models that operate under the principles of competition. The resulting\nneural system is shown to offer an effective approach for combating\ncatastrophic forgetting in online continual classification problems. We\ndemonstrate that the proposed CCM system not only outperforms other competitive\nlearning neural models but also yields performance that is competitive with\nseveral modern, state-of-the-art lifelong learning approaches on benchmarks\nsuch as Split MNIST and Split NotMNIST. CCM yields a promising path forward for\nacquiring representations that are robust to interference from data streams,\nespecially when the task is unknown to the model and must be inferred without\nexternal guidance.\n