Historical linguists who specialize in Japanese agree that it is one of the two members of the Japonic language family, but remain divided as to the origins of the Japonic languages. An older view, still widely held by some linguists and many non-linguists, is that Japanese is a language isolate.

As for its relation to other languages, there are several theories (presented roughly in descending order of certainty):

• Japanese is a member of the Altaic language family. Other languages in this group include Mongolian, Tungusic, Turkish, and (according to most proponents) Korean. Evidence for this theory lies in the fact that like Turkish and Korean, Japanese is an agglutinative language. Additionally, there are a suggestive number of apparently regular correspondences in basic vocabulary, such as ishi "stone" to Turkic daş, yon "four" to Turkic dört, kura "saddle" to Turkic kürtün, kiru "to cut" to Turkic kir-, inu "dog" to Turkic it, kuro "black" and kurai "dark" to Turkic köl "shadow", etc.
• Japanese is a relative of extinct languages spoken by historic cultures in what are now the Korean peninsula and Manchuria. The best attested of these is the language of Goguryeo (a.k.a. Koguryo), with the more poorly-attested languages of Baekje (a.k.a. Paekche) and Buyeo (a.k.a Puyo) hypothesized to also be related. The limited data on these languages, as well as these cultures' historic ties, are the primary evidence. This has been largely subsumed into the Altaic theory.
• Japanese is a relative of Korean. This theory is based on the high degree of similarity between Japanese and Korean grammar. Proponents of this theory have also proposed Japanese-Korean cognates. The idea of a Japanese-Korean relationship has been largely subsumed into the Altaic theory.
• Japanese is a creole language. Phonological similarities and geographical proximity to Austronesian languages have led to the theory that Japanese may be a kind of creole, with an Altaic substratum and an Austronesian superstratum, or vice versa.
• Japanese is a purely Austronesian language. This theory enjoys little currency, since the grammar and lexis of Japanese are vastly different from those of any known Austronesian language.
• Ono Susumu has suggested a possible relationship between Japanese and Tamil, a member of the Dravidian language family spoken in southern India.

Specialists in Japanese historical linguistics all agree that Japanese is related to the Ryukyuan languages (including Okinawan); together, Japanese and Ryukyuan are grouped in the Japonic languages. Among these specialists, the possibility of a genetic relation to Goguryeo has the most evidence; relationship to Korean is considered plausible but is still up to debate; the Altaic hypothesis has somewhat less currency, though it has grown significantly more respectable in recent years, primarily due to the work of Sergei Starostin, et al. Almost all specialists reject the idea that Japanese could be genetically related to Austronesian/Malayo-Polynesian languages or Sino-Tibetan languages, and the idea that Japanese could be related to Tamil is almost entirely excluded.

It should be noted that linguistic studies, like all fields, can be strongly affected by national politics and other non-academic factors. For example, most linguists would say that Romanian and Moldovan are essentially the same language, and that they are known as two different languages for political reasons. Japan's long-standing rivalries and enmities with virtually all of its neighbours make the study of linguistic connection particularly fraught with such political tensions. However, these tensions are less prevalent among non-Japanese researchers.