Lusitanian (so named after the Lusitani or Lusitanians) was an Indo-European Paleohispanic language. There has been support for either a connection with the ancient Italic languages or Celtic languages. It is known from only six sizeable inscriptions, dated from c. 1 AD, and numerous names of places (toponyms) and of gods (theonyms). The language was spoken in the territory inhabited by Lusitanian tribes, between the Douro and Tagus rivers, territory that today falls in central Portugal and western Spain.

In spite of the limited corpus, the Lusitanian inscriptions are linked by several recurring words, which enabled the establishment of Lusitanian as a linguistic entity. The number of semantically clear words is minimal. Three words for sacrificial animals are clearly identified (porcom pig, pork, taurom bull, oilam sheep, lamb).

Lusitanian has undergone the same centum development as all Western IE languages (e.g. porcom from *pork̑om) and u̯ did not merge, but rather that *k u̯ became om). Furthermore, evidence for the fate of labiovelars is arguable. It has been suggested that unlike in other kentum-languages PIE *k u̯ and *ku̯ /k ̑ u̯ did not merge, but rather that *k u̯ became Lus. p and *ku̯ /k ̑ u̯ remained as Lus. <qu> (Prósper 2002:396−397; Witczak 2005: 274−276). s generally remains in Lusitanian but is dropped on occasion in final position (Stifter 2010− 2011: 189−190).

Scholars like Untermann (1987) identify toponymic and anthroponymic radicals which are clearly linked to Celtic materials: briga ‘hill, fortification’, bormano ‘thermal’ (Cf. theonym Bormo), karno ‘cairn’, krouk ‘hillock, mound’, crougia ‘monument, stone altar’, etc. Others, like Anderson (1987), point to results of inscriptional comparisons between Lusitania and Gallaecia that they argue show, somewhat indirectly, that Lusitanian and Gallaecian formed a fairly homogeneous linguistic group. Indigenous divine names in Portugal and Galicia frequently revolve around the gods or goddesses Bandu (or Bandi), Cossu, Nabia and Reve:

The Lusitanian and Gallaecian divine name Lucubos, for example, also occurs outside the peninsula, in the plural, in Celtic Helvetia, where the nominative form is Lugoves. Lug was also an Irish god, and the ancient name of Lyon was Lug dunum and may have a connection with the Lusitanian and Gallaecian word, suggesting therefore a north-western Iberian sprachbund with Lusitanian as a dialect, not a language isolate. Ellis Evans believes that Gallaecian-Lusitanian were one same language (not separate languages) of the “P” Celtic variant.

While chronology, migrations and diffusion of Hispanic Indo-European peoples are still far from clear, it has been argued there is a case for assuming a shared Celtic dialect for ancient Portugal and Galicia-Asturias. Linguistic similarities between these Western Iberian Indo-Europeans, the Celtiberians, the Gauls and the Celtic peoples of Great Britain indicate an affiliation in vocabulary and linguistic structure.

Furthermore, scholars such as Koch say there is no unambiguous example of the reflexes of the Indo-European syllabic resonants *l̥, *r̥, *m̥, *n̥ and the voiced aspirate stops *bʱ, *dʱ, *ɡʱ. Additionally, names in the inscriptions can be read as undoubtedly Celtic, such as AMBATVS, CAELOBRIGOI and VENDICVS. Dagmar Wodtko argues that it is hard to identify Lusitanian personal or place-names that are actually not Celtic. These arguments contradict the hypothesis that the p- in PORCOM alone excludes Lusitanian from the Celtic group of pre-Roman languages of Europe and that it can be classed as a Celtic dialect but one that preserved Indo-European *p (or possibly an already phonetically weakened [ɸ], written P as an archaism). This is based largely on numerous Celtic personal, deity, and place names.

Lusitanian possibly shows /p/ from Indo-European *kʷ in PVMPI, pronominal PVPPID from *kʷodkʷid, and PETRANIOI derived from *kʷetwor- 'four', but that is a feature found in many Indo-European languages from various branches (including P-Celtic/Gaulish, Osco-Umbrian, and partially in Germanic where it subsequently evolved into /f/, but notably not Celtiberian), and by itself, it has no bearing on the question of whether Lusitanian is Celtic. Bua Carballo suggests that pairings on different inscriptions such as Proeneiaeco and Proinei versus Broeneiae, and Lapoena versus Laboena, may cast doubt on the presence of a P sound in Lusitanian.