Peter Blouw

Knowledge and luck

Nearly all success is due to some mix of ability and luck. But some successes we attribute to the agent’s ability, whereas others we attribute to luck. To better understand the criteria distinguishing credit from luck, we conducted a series of four studies on knowledge attributions. Knowledge is an achievement that involves reaching the truth. But many factors affecting the truth are beyond our control, and reaching the truth is often partly due to luck. Which sorts of luck are compatible with knowledge? We found that knowledge attributions are highly sensitive to lucky events that change the explanation for why a belief is true. By contrast, knowledge attributions are surprisingly insensitive to lucky events that threaten, but ultimately fail to change the explanation for why a belief is true. These results shed light on our concept of knowledge, help explain apparent inconsistencies in prior work on knowledge attributions, and constitute progress toward a general understanding of the relation between success and luck.

Using Neural Networks to Generate Inferential Roles for Natural Language

Neural networks have long been used to study linguistic phenomena spanning the domains of phonology, morphology, syntax, and semantics. Of these domains, semantics is somewhat unique in that there is little clarity concerning what a model needs to be able to do in order to provide an account of how the meanings of complex linguistic expressions, such as sentences, are understood. We argue that one thing such models need to be able to do is generate predictions about which further sentences are likely to follow from a given sentence; these define the sentences “inferential role.” We then show that it is possible to train a tree-structured neural network model to generate very simple examples of such inferential roles using the recently released Stanford Natural Language Inference (SNLI) dataset. On an empirical front, we evaluate the performance of this model by reporting entailment prediction accuracies on a set of test sentences not present in the training data. We also report the results of a simple study that compares human plausibility ratings for both human-generated and model-generated entailments for a random selection of sentences in this test set. On a more theoretical front, we argue in favor of a revision to some common assumptions about semantics: understanding a linguistic expression is not only a matter of mapping it onto a representation that somehow constitutes its meaning; rather, understanding a linguistic expression is mainly a matter of being able to draw certain inferences. Inference should accordingly be at the core of any model of semantic cognition.