Filipino shamans, commonly known as babaylan (also balian or katalonan, among many other names), were shamans of the various ethnic groups of the pre-colonial Philippine islands. These shamans specialized in communicating, appeasing, or harnessing the spirits of the dead and the spirits of nature. They were almost always women or feminized men (asog or bayok). They were believed to have spirit guides, by which they could contact and interact with the spirits and deities (anito or diwata) and the spirit world. Their primary role were as mediums during pag-anito séance rituals. There were also various subtypes of babaylan specializing in the arts of healing and herbalism, divination, and sorcery.
The most common native terms for shamans among Austronesian groups in Island Southeast Asia are balian, baylan, or cognates and spelling variants thereof. They are all derived from Proto-Western-Malayo-Polynesian *balian, meaning "shaman" (probably originally female, transvestite, or hermaphroditic) or "medium". Various cognates in other non-Filipino Austronesian languages include babalian, bobolian, and bobohizan (Kadazan-Dusun); wadian (Ma'anyan); belian (Iban); belian (Malay); walen or walyan (Old Javanese); balian (Balinese); bolian (Mongondow); balia (Uma); wulia or balia (Bare'e); balia (Wolio); balian (Ngaju); and balieng (Makassar). However *balian-derived terms have largely disappeared among lowland Filipinos after Christianization in the Spanish era. Some exceptions include Bikol where it persisted and acquired the Spanish feminine suffix -a as balyana. It also survives among some Muslim Filipinos like in Maranao walian, although the meaning has shifted after Islamization.
The linguist Otto Dempwolff has also theorized that *balian may have ultimately derived from Proto-Austronesian *bali ("escort", "accompany") with the suffix *-an, in the meaning of "one who escorts a soul to the other world (a psychopomp)". However, the linguists Robert Blust and Stephen Trussel have noted that there is no evidence that *balian is a suffixed form, and thus believe that Dempwolff's interpretation is incorrect.