I had intended, on an essay on this subject, to set out to show that the Pictish Language was a form of P-Celtic like Welsh, Cornish and Breton. However, since it seems that linguists nowadays agree that that was the case, there would be little point in rehashing their arguments.
For background information relevant to this story, let it be understood that the Celtic languages are conventionally divided into P-Celtic and Q-Celtic, the latter consisting of the Gaelic languages of which Irish is the most important representative. A theory currently in favour is that early Indo-European had an uvular stop like the qoph in Hebrew or the qof in Arabic. This sound mutated into the unvoiced stops P, K and T depending on the phonemic environment, in such a way that some languages can be characterised by their preference for one over the others.
For example, where Greek has hippos “horse”, Latin has equus. Greek is therefore a P language while Latin is a Q language. Compare Latin quo “where” with Greek pou; Irish cé “what” with Welsh pa; Latin quinque and Irish cuig “five” with Greek pente and Welsh pump. Sometimes foreign borrowings get changed accordingly: take Greek Paskha “Easter”, Latin Pascha, French Pâques, Welsh Pasg, but Irish Cáisg.
Any reflection on the Pictish language must take into account the explicit statement by Bede (672/3 – 735) that Pictish was a different language from Welsh. The four languages that were in use in Britain in his day were English (in various dialects), British (Welsh), Irish (Gaelic) and Pictish. Bede, being based in Monkwearmouth, spoke the first three, as well as a particularly limpid Latin, because of his pivotal position in the propagation of Christianity. It would not be surprising if he also knew Pictish.
All we know of Pictish, since there are no written texts, is derived from place names and one or two personal names. The Pictish kingdom stretched from the northern shore of the Firth of Forth (“Ffwrdd Ffrwydd” in Welsh), to as far north as perhaps Sutherland, along the eastern side of modern Scotland. The place names in this region are a splendid muddle. On the same coastline we have Aberdeen (Welsh aber = “estuary”) and Inverbervie (Irish inbhear = “estuary”). A lot of names are clearly similar to Welsh; though many place names have been hibernicised and anglicised and subjected to idiosyncratic transliteration until they just look odd. The one marker that we have for working out the extent of Pictland is the presence of place names beginning “Pit”.
Pictland disappeared as a political entity when Kenneth MacAlpin (840-858) united it with his own kingdom to create the basis of a united Scotland. It is interesting that he named his daughter “Buthuc” (modern Welsh buddug = Victory or Victoria). Presumably he was intending to give her a Pictish name to cement the unification of the two kingdoms. The Pictish language went on to suffer from encroachment from all sides: Gaelic from the west, English from the south, and Norse from the north and from the sea. It must also have suffered from not being written. One commentator suggests that illiteracy was a deliberate policy. Perhaps the Picts thought, like Boko Haram in present-day Nigeria, that ignorance and illiteracy were a protection against external interference.
Now we come to the only original contribution to this discussion that I think I can make. It comes down to the word “Pit”, which does not seem to be a Welsh word and which was therefore used as evidence at one time to suggest that the Pictish language was non-Indo-European. The name that first put me on to this quest was the surname “Pitblatho”, belonging to one of Churchill's secretaries during the Second World War. “Blatho” seemed so much like the modern Welsh bleiddiau (= “wolves”) that I checked and found that Pitblatho was a settlement on the outskirts of Aberdeen. The surmise was therefore that the name was the equivalent of Wolverhampton or Wolvercote or one of the numerous Wolvertons. So “Pit” would seem to mean some kind of settlement. I tried other Pits. My most interesting discovery was that Pittenweem, on the south coast of what is still called “The Kingdom of Fife”, could be readily interpreted as Pit yng ngwymon (= “Pit among the seaweed”); that fits the geography. Most Pits proved incomprehensible, though, either because of my lack of Welsh or because the second part of the word was probably a personal name on the analogy of “Kensington”.
Then I started to wonder. If Pictish was truly P-Celtic, could not Pit derive from an alien word Kit? What else could this be but the Latin civitas (Italian città, French cité, Spanish ciudad, Occitan ciotat with a hard “c”, Portuguese cidade, English city)? A Roman city was an organised urban structure with a defensive wall and a municipal government. Modern Welsh does not have this word; their word for “city” is dinas, referring to the fortifications, not the system of government. My theory is that some king of Pictland, desirous of adopting some of the trappings of Roman civilisation, bestowed on some favoured communities the privilege of a municipal council. Perhaps these were the first Scottish burghs. How can it be proved?
As for Bede's assertion, let us suppose that Pictish stood in the same relationship with Welsh as Spanish to Portuguese, Dutch to German, Xhosa to Zulu. Then we are not challenging Bede.