The formal framework introduced by Kratzer in 1981 is closely related to the work on DSDL in the sixties and seventies, such as the work of Hansson and Lewis discussed in the handout of Lecture 1. Her semantics is therefore sometimes referred to as the Lewis-Kratzer semantics. In the semantics of Kratzer, the preference ordering is not given, but it is constructed from two sets of propositions,...

Many parsers learn sparse class distributions over trees to model natural language. Recursive Neural Networks (RNN) use much denser representations, yet can still achieve an F-score of 92.06% for right binarized sentences up to 15 words long. We examine an RNN model by comparing it with an abstract generative probabilistic model using a Deep Belief Network (DBN). The DBN provides both an upward...

External negation of conditionals occurs in sentences beginning with ‘It is not true that if’ or similar phrases, and it rare natural language. A conditional may also be denied by another the same antecedent opposite consequent. Most often, when implicative, denying one concessive, vice versa. Here I argue that, language pragmatics, ‘If $A$, $\sim B$’ entails ‘$\sim$(if $A, B$)’, but B$)’ does ...

We prove that continuous sentences preserved by the ultramean construction (a generalization of the ultraproduct construction) are exactly those sentences which are approximated by linear sentences. Continuous sentences preserved by linear elementary equivalence are exactly those sentences which are approximated in the Riesz space generated by linear sentences. Also, characterizations for linea...

This paper begins with a review of the literature on plausible reasoning with deductive arguments containing a conditional premise. There is concurring evidence that people presented with valid conditional arguments such as Modus Ponens and Modus Tollens generally do not endorse the conclusion, but rather find it uncertain, in case (1) the plausibility of the major conditional premise is debata...

In this paper, we present some recent improvements in our automatic speech segmentation system, which only needs the speech signal and the phonetic sequence of each sentence of a corpus to be trained. It estimates a GMM by using all the sentences of the training subcorpus, where each Gaussian distribution represents an acoustic class, which probability densities are combined with a set of condi...

This article presents axioms for comparative conditional probability relations. The axioms presented here are more general than usual. Each comparative relation is a weak partial order on pairs of sentences (i.e. each relation is transitive and reflexive) but need not be a complete order relation. The axioms for these comparative relations are probabilistically sound for the broad class of cond...

We describe a simple improvement to ngram language models where we estimate the distribution over closed-class (function) words separately from the conditional distribution of open-class words given function words. In English, function words account for about 30% of written language, and also form a natural skeleton for most sentences. By factoring a language model into a function word model an...