Damning (adjective) /ˈdæmnɪŋ/

It’s a phonotactic rule that two nasals cannot occur next to each other within a tautosyllabic* cluster. Because /m/ and /n/ are both nasals, the tautosyllabic cluster /mn/ is therefore forbidden in English.

        * ^{Tautosyllabic means within the same syllable.}

That’s one of the reasons we get a silent n in damn.

In some cases, the /n/ gets pronounced when a vowel-initial suffix follows the ‹mn›, though the pattern is very irregular. See for example damnable pronounced [ˈdæm.nəbl̩] — the vowel in ‑able split up the cluster /mn/. The same thing happens in damnation. Damning on the other hand is not pronounced with an /n/. However, some sources claim that the pronunciation with an /n/ has existed.

The only source I could find was the Modern English Grammar on Historical Principles (7 vols.). 1–7 (1949) by Otto Jespersen. The entry on /mn/ reads:

After /m/, a final /n/ has been lost (assimilated to /m/): damn /damn/ now [dæm], condemn, hymn, limn, column, solemn, autumn. The loss is shown by such inverse spellings as solembe, (Sh LL V. 2.118, quarto of 1598); C 1627 expressly says that n is mute in solemne and hymne. N has been everywhere retained in spelling, except in occasional dam (for damned?: Meredith EH 134 “and dam rum chaps they were!”).

Homonyms: damn = dam, hymn = him, limn = limb formerly lim.

Before a vowel, /n/ is retained: damnation, condemnation, damnable, autumnal, solemnity. Before ‑ing /n/ was formerly heard in “the solemn articulation of damning, condemning etc.” (E 1766, also Walker); now the pronunciation without [n] has been analogously extended to these forms, though the NED has both pronunciations for the participle (but not the verbal noun) damning and recognizes [ˈdæmnɪd] as a poetical form of damned by the side of [dæmd].

I don’t think I can add anything to this explanation.

Also, if you’ve noticed, most adjectives that end with ‑ed have final [ɪd] in their pronunciation (e.g. wicked, naked, learned). In my opinion, damned followed the same sequence at some point ([ˈdæmnɪd]).

I think the shift from amn’t to ain’t is also relevant. The contraction of am not used to be amn’t, where we see the same cluster /mn/, but it got reduced to /n/. Here it’s /n/ and not /m/ because of the following /t/, and also per Phonotactics^{[citation needed]} because:

a nasal following an obstruent in the coda should be homorganic with the obstruent.