How is it that words (such as “Aristotle”) come to stand for the things they stand for (such as Aristotle)? Is the thing that a word stands for, its reference, fully identified or described by conventions known to the users of the word? Or is there a more roundabout relation between the reference of a word and the conventions that determine or fix it? Do words like “water,” “three,” and “red” refer to appropriate things, just as the word “Aristotle” refers to Aristotle? If so, which things are these, and how do they come to be referred to by those words? In Roads to Reference, Mario Gómez-Torrente provides novel answers to these and other questions that have been of traditional interest in the theory of reference. The book introduces a number of cases of apparent indeterminacy of reference for proper names, demonstratives, and natural kind terms, which suggest that reference-fixing conventions for them adopt the form of lists of merely sufficient conditions for reference and reference failure. Arguments are then provided for a new anti-descriptivist picture of those kinds of words, according to which the reference-fixing conventions for them do not describe their reference. The book also defends realist and objectivist accounts of the reference of ordinary natural kind nouns, numerals, and adjectives for sensible qualities. According to these accounts, these words refer, respectively, to “ordinary kinds,” cardinality properties, and properties of membership in intervals of sensible dimensions, and these things are fixed in subtle ways by associated reference-fixing conventions.