In this article we present a syntax-based translation system, called
TABL (Translation using Alignment-Based Learning).  It translates
natural language sentences by mapping grammar rules of the source
language to those of the target language.  All grammar rules are
induced using the Alignment-Based Learning grammatical inference
framework.  By parsing a sentence in the source language, the grammar
rules that are used in the derivation are translated using the mapping
and subsequently, a derivation in the target language is generated.
Here, we illustrate that this is a valid machine translation approach,
as the initial results are encouraging.  We assess some remaining
problems and current restrictions to this approach and propose
possible solutions.