In Japanese, as in many other head-final languages, it is often hard to impossible to distinguish between clitics and suffixes. -de can only attach to a noun be

In Japanese, as in many other head-final languages, it is often hard to impossible to distinguish between clitics and suffixes. -de can only attach to a noun because a noun phrase always ends with a noun and there can be nothing between a noun and a -de, they are inseparable. On the other hand, there is no phonological fusion of any kind between a noun and -de (unlike in verbs, where yom-u 'to read' becomes yon-de in the gerund form). Because of this -de and other case markers are described both as "particles" and "case endings" in different sources. In agglutinative languages with freer word order, it is easier to argue that a locative morpheme is a suffix because it always appears with nouns, but absent any phonological processes that bind them together, one can always write them with a space and call the second part "locative postposition". In some cases, it is even unclear if an agglutinative language is very morphologically complex or just uses a lot of function words.