# Equiprobable mappings in weighted constraint grammars

## TL;DR

This paper compares MaxEnt and stochastic HG frameworks in weighted constraint grammars, showing MaxEnt's ability to distinguish all mappings and characterizing when stochastic HG produces equiprobable mappings, with a case study on Finnish stress.

## Contribution

It provides a formal characterization of equiprobable mappings in stochastic HG and contrasts this with MaxEnt's discriminative power.

## Key findings

- MaxEnt can distinguish any two different mappings.
- Stochastic HG admits equiprobable mappings, characterized formally.
- Comparison on Finnish stress illustrates differences in predictions.

## Abstract

We show that MaxEnt is so rich that it can distinguish between any two different mappings: there always exists a nonnegative weight vector which assigns them different MaxEnt probabilities. Stochastic HG instead does admit equiprobable mappings and we give a complete formal characterization of them. We compare these different predictions of the two frameworks on a test case of Finnish stress.