Editing Openai/69186d20-2d14-8013-b9c6-c09000f173f6 (section)

=== Assistant: 2511141650ET ===
251114_1650ET

short answer: they’re cousins, not direct equivalents. Old Irish fortis/lenis in sonorants and Welsh voiceless sonorants both grow out of Celtic gemination/fortition, but Old Irish ll was not Welsh-style [ɬ], and Welsh ll didn’t “come from” Old Irish — they’re separate branch outcomes of a shared Proto-Celtic tendency.

===== Old Irish had four-way coronal sonorant contrasts: =====
* broad vs slender
* fortis vs lenis

For l, n, r, the fortis ones were written ll, nn, rr; lenis as l, n, r.

Phonetically (best current view):
* Fortis sonorants were: - tenser, often longer, with more tongue contact.
* Lenis sonorants were: - shorter, laxer, sometimes closer to nasalized glides in certain contexts.

Crucially:
* They were still voiced sonorants, not full-blown voiceless fricatives like Welsh [ɬ].
* The contrast was robust between vowels and word-finally, e.g.: - coll “hazel” vs col “sin” - sonn “stake” vs son “sound”

Over time:
* True length/gemination was lost,
* ll, nn, rr were repurposed as “fortis” spellings,
* In most modern Irish dialects the fortis/lenis contrast has largely collapsed, with only broad vs slender really active; the doubles are mostly orthographic fossils, though some dialects still have a 4-way coronal system.

So: Old Irish ll ≈ tense/fortis [l̪ː] or [L], voiced, not [ɬ].

===== Modern Welsh has genuinely voiceless sonorants, most famously: =====
* ll = [ɬ], voiceless alveolar lateral fricative
* rh = voiceless rhotic [r̥]
* plus mh, nh, ngh for voiceless nasals in mutation contexts

These arose in Brythonic from internal developments, e.g.:
* clusters with /h/ and liquids/nasals → devoicing,
* mutation systems that created environments for voiceless sonorants, later phonologised.

So Welsh ll is a bona fide phoneme [ɬ], and orthographic ll is dedicated to that.

===== In a deep, structural sense: =====
* Proto-Celtic developed a strong tendency to: - create geminate sonorants, and - allow fortis vs lenis contrasts to carry morphological weight.

From that common base:
* Goidelic branch (Old Irish → modern Irish/Scottish/Manx) - Geminate sonorants > fortis vs lenis distinctions (tense vs lax, not voiceless vs voiced). - Later mostly collapsed; double letters survive as historical markers.
* Brythonic branch (Welsh/Breton/Cornish) - Some sonorants, especially laterals and rhotics, drifted toward devoicing in certain contexts. - In Welsh this devoicing phonologised as ll = [ɬ], rh = [r̥], etc.

So the relationship is:

: 

Not:

: 

Different branches, parallel but independent exploitations of the same inherited “we like beefed-up sonorants” tendency.

===== - As a distinct phonetic category (length / strong vs weak): - Mostly yes in modern Irish: the fortis–lenis contrast is no longer a clear, productive opposition. =====
* As orthography / residual tensity: - Partly no: - You still see ll, nn, rr in spelling, - some dialects (esp. conservative Irish/Scottish) still have noticeably tens(er) versions of those consonants in certain positions.

But none of that is a direct ancestor of Welsh [ɬ]. That voiceless lateral is a Brythonic innovation from the same wider Celtic gemination/fortition background, not a handoff from Old Irish.