In Upper Chinook [...], 't:d' exists objectively but not psychologically; one says, e.g., 'inat' 'across', but 'inad' before words beginning with a vowel, and the two forms of the final consonant are [...] felt to be the 'same' sound [...]. The Upper Chinook 'd' exists only as a mechanical variant of 't' [...] - Sapir (1925), a pag.38