Notes on Classical Gaelic Grammar

From Celtic Languages
Jump to navigationJump to search

Classical Gaelic or Classical Irish (Gáoidhealg or Gaoidhealg /ɡəːð´əlɡ ~ ɡəːɣ´əlɡ/ in the language itself) was the standard language of bardic poetry in Ireland and Scotland between 13th and 18th century. It was created in late 12th century by poets trying to bring the poetic language closer to the vernacular of that time, the standard allowed a lot of spoken forms used in the dialects of late 12th century, even though it was itself highly prescriptive.[gr 1]

While the spoken language evolved, the grammar of the language used in poetry was kept fixed throughout centuries (although new vocabulary was used in poetry). The spoken language of 13th–18th century Ireland (and one used in prose texts) is called Early Modern Irish while the term Classical Gaelic is typically reserved for the standardized conservative language of the bardic poetry.

This article deals with elements of grammar of this bardic standard (although it might mention some developments rejected in the bardic standard occasionally). But bear in mind it’s more of a random collection of notes (mostly from Bardic Syntactical Tracts[gr 2] and IGT i[gr 3][gr 4]) than a comprehensive grammar. It’s also a work in progress.


Early Modern Irish and Scottish manuscripts vary a lot in the spelling they used. The earliest manuscripts use basically the Middle Irish spelling that’s closer to Old Irish system than to modern orthography, later ones start to write consonants in a more modern way and mark lenition (semi-)consistently using ponc séimhithe (dot placed above the letter, sometimes a little h instead of a dot) but there are several ways to mark eclipsis (namely, the letters t and c are often doubled to tt, cc in eclipsing context). [TODO: expand]

Modern normalized editions of Classical Gaelic and Early Modern Irish texts often use the modern pre-reform Irish spelling as one can find eg. in Dinneen’s dictionary. But that’s often not the spelling used in manuscripts – and one won’t find it often in dyplomatic editions of classical texts.

Among the common differences between the spellings used commonly in manuscripts and the modern normalized spelling are:

  1. the clusters /st, sk, sp/ are written sd, sg, sb (this varies in normalized texts), eg. innsgne, subsdainnteach vs inscne, substainteach,
  2. the ‘long diphthongs’ and the ao vowel mark their length with a ‘fada’: Gáoidhealg, táobh, síad, íasg, búadh vs modern Gaoidhealg, taobh, siad, iasc/iasg, buadh,
  3. the long é before a broad consonant stands on its own, breaking the leathan le leathan rule (while it is the expected éi before a slender one): dénamh, bél, slégar but Éire vs déanamh, béal, sléagar, Éire,
  4. eclipsis of t and c written as tt, cc (and pp for p?): a ttigh ‘in a house’ vs i dtigh / a dtigh.

This page follows the spelling conventions of Mac Cárthaigh edition of Irish Grammatical Tracts I (“The Art of Bardic Poetry”, ABP) [gr 3], that is, unlike most modern editions, it normalizes the spelling to form closer to the older usage described above.

[TODO: something about spelling of reduced unstressed vowels: a vs u vs o; i vs ei]



Classical Gaelic nouns and adjectives were declined in 5 distinct grammatical cases:

  • nominative (ainm, lit. ‘name’) – the grammatical subject of the verb and also predicative of copular clauses,
  • accusative (often réim, lit. ‘succession’, the term is used generally for ‘an oblique case’, accusative plural is often called tochlaghadh, tochlughadh) – the direct object of a verb, also complement of some prepositions, especially when movement is involved,
  • genitive (táoibhréim, lit. ‘side-case, by-case’) – expresses possessor, or otherwise connects two nouns (like in Modern Irish or Scottish Gaelic),
  • dative (tuilréim, tuillréim) – used for complements of most prepositions (so perhaps prepositional would be a better name),
  • vocative (sometimes agallaimh, gairm agallmha, ‘address, addressing call’, typically not recognized as separate case in classical grammars) – used when directly addressing someone or something.

The case system of Old Irish (which itself lost most of case endings due to apocope) got somewhat eroded during Middle Irish times because of the loss of distinction between unstressed /ə/ and /u/, and also merging of all unstressed absolute final vowels. Nevertheless all the Old Irish cases and their uses survived til Classical Gaelic times even if their forms merged in some noun classes.

Classical grammars recognize two grammatical genders: masculine (feirinnsgne) and feminine (baininnsgne), with the nouns whose accusative singular is the same as nominative singular (ie. mostly nouns declining according to old o-stem pattern) commonly referred to as ‘masculine’, and those whose acc. sg. = dat. sg. (mostly those declining acc. to old ā-stem pattern and consonant stems) referred to as ‘feminine’ (even if actually belonging to the other gender).

Réim connsaine

When a noun was a direct object of the verb it generally had to be in the accusative case and underwent lenition. This lenition is called réim connsaine (lit. ‘consonant accusative’ or ‘consonant case’). If the noun declined according to masculine pattern, ie. it had its acc. sg. form equal to nom. sg., then réim connsaine was optional and nominative could be used instead. When one used a noun with a ‘feminine’ declension (ie. the acc. sg. differed from nom. sg.), then both lenition and accusative form were obligatory:

  • do-chiú chnáoi ngil ‘I see a bright nut’ – here the lenited accusative form chnáoi must be used since it differs from nominative cnú, it also must eclipse its attributes, since accusative case requires eclipsis,
  • ní fhaca choin ‘I did not see a dog’, ní fhaca bhoin ‘I did not see a cow’ – the acc. sg. forms differ from nom. ‘a dog’ and ‘a cow’,
  • do bhúaileas choin ngil ngirr ccáoich ‘I hit/stroke a blind short bright dog’,


  • do-chiú shúil ngil or do-chiú súil gheal ‘I see a bright eye’, because acc. sg. súil is the same as nominative,
  • do mharbhas fhear mbeag or do mharbhas fear beag ‘I killed a small man’ – because fear is both nominative and accusative form, either can be used; the only requirement being that if the direct object is lenited (has réim connsaine), then it must eclipse its attribute, because it is visibly accusative.

Since many nouns have accusative plural (tochlaghadh) distinct from their nominative plural, they also are lenited:

  • do-chiú fhiora beaga ‘I see small men’ – acc. pl. distinct from nom. pl. fir bheaga,
  • also note eg. dún do liobhra ‘close your books’ – acc. pl. distinct from nom. pl. leabhair (even though lenition is not visible in writing, and it’s required here by possessive do ‘your’).

The accusative singular definite article eclipses:

  • gearr an ccráoibh ‘cut the branch’ (nom. sg. an chráobh),
  • muirfead an bhfilidh ‘I will kill the poet’ (nom. sg. an file)

but the eclipsis is optional in direct objects if the form is the same as nominative (a similar rule to réim connsaine):

  • gonfad an bhfear or gonfad an fear ‘I will wound the man’.

[TODO: other situations where acc. form is suppressed – infix pronouns, use of verbal noun phrases, dir. obj. after acht, nó, etc.]


Classical Gaelic had a peculiar rule regarding attributed nouns in the genitive case. The rule is called slégar (sléagar in modern spelling, sometimes also slégur, sléagur). As explained in Mac Cárthaigh’s ABP[gr 3], pp. 295–298, see also his article[gr 5], slégar can be applied in phrases that have structure like this:

beginning middle end
noun noun (or two) in the genitive noun in the genitive
or adjective
or sloinneadh (surname in the genitive form, like mheic Dhíarmada or Í Bhríain)

In phrases like that one can:

  • change eclipsis or lack of lenition in the middle part into lenition,
  • or remove lenition in the middle part.

Thus the regular grammatical way to state ‘with the wife of Brian mac Taidhg’ would be:

  • le mnáoi mBríain mheic Thaidhg (le takes accusative, thus the genitive Bríain is regularly eclipsed)

but with slégar applied it can also be:

  • le mnáoi Bhríain mheic Thaidhg (with the eclipsis changed into lenition).

And ‘bright Brian’s spear’ would regularly be:

  • sleagh Bhríain ghil (sleagh ‘spear’ is a feminine noun, thus it lenites its attribute),

but with slégar applied it also can be:

  • sleagh Bríain ghil.

Slégar is applicable both to definite and indefinite phrases, for example:

  • adhaltras fhear na cruinne ‘the adultery of the men of the world’ is an example of slégar because adhaltras ‘adultery’ is a masculine noun and thus should not lenite its attribute,
  • the phrase would regularly be: adhaltras fear na cruinne.

There are situation where slégar is prevented. One of them is that an eclipsed noun in genitive blocks slégar in subsequent nouns:

  • ré néimh bhfáinne fir óir is the regular way to say ‘to brightness of a golden man’s ring’ (bhfáinne ‘of a ring’ is an io-stem masculine noun in genitive, it allows its attribute – fir ‘a man’s’, in genitive – to stay unlenited[1]),
  • ré néimh fháinne fhir óir and ré néimh fháinne fir óir are correct with slégar applied (eclipsis turned lenition on bhfáinne, and also possible slégar on fir),
  • but re néimh bhfáinne fhir óir is incorrect (because bhfáinne is eclipsed, the subsequent fir must stay unlenited and slégar cannot be applied to it).

Another such case is genitives connected with agus ‘and’, ‘or’, or ‘nor’: they block slégar in the sense that they connect two unattributed genitives, they break genitive chain, ie. mac fir agus mná means ‘a son of a man and of a woman’, none of the genitives is attributing the other one, so slégar cannot be applied to either (but slégur may be applied before the conjunction if the first genitive is attributed, eg. mac fir óig agus mná without slégar and mac fhir óig agus mná with slégar are both correct).

It’s worth noting that the operation of slégar, in cases where it causes regularly unexpected lenition, is very reminiscent of Modern Irish ‘functional genitive’ – lenition of definite nominative nouns replacing the genitive form (for example ‘Cáit’s house’ being teach Cháit, or ‘the fisher’s boat’s mast’ being crann bhád an iascaire) – but note the differences:

  • unlike in Modern Irish, slégar does not prevent the genitive form – all the nouns in a chain in Classical Gaelic are in genitive, note the adhaltras fhear na cruinne example above with gen. pl. fear ‘of men’, in Modern Irish it’d be adhaltras fhir na cruinne,
  • slégar affects both definite and indefinite phrases,
  • slégar can also undo lenition, changing regularly lenited noun into an unmutated one, and it can change eclipsis into lenition.


Stationary dative vs accusative of motion

Classical Gaelic keeps the distinction between accusative and dative grammatical case used after certain prepositions. Accusative is used to convey the motion (eg. i + accusative means ‘into’) while dative expresses stationary position (eg. i + dative means ‘inside’). It’s the same distinction as can be found in German, Latin (with stationary ablative), Slavic languages (with stationary locative).

[TODO: list all prepositions taking acc/dat depending on meaning]

For example:

  • Cuirfead mo láimh ar fear or ar fhear ‘I will put my hand upon a man’ with accusative (motion),
  • Bíaidh mo lámh ar fior ‘my hand will be on a man’ with dative (stationarily placed).

But some prepositions take just one of the cases, eg. ós always takes the dative:

  • Cuirfead mo láimh ós fhior ‘I will put my hand above a man’ with dative (even though motion is involved here).

There are also compound “prepositions” that take genitive – they are generally phrases with the second element being a noun which itself is either in dative or accusative, and thus they themselves cause either lenition (if the noun-part is in dative) or eclipsis (if the noun-part is in accusative). Some of them change the mutation they cause based on their meaning:

  • tar éis fhir or ar éis fhir ‘after a man’ (lit. ‘on, over a track of a man’) – in the meaning ‘after’ éis is in dative and causes lenition,
  • tar éis bhfir or ar éis bhfir ‘instead of a man’ (lit. ‘onto, over a track of a man’) – in the meaning ‘instead of’ éis is in accusative and causes eclipsis.

But some of them cause always the same mutation regardless of meaning:

  • d’éis fhir can mean both ‘after a man’ and ‘instead of a man’ (do can be used instead of ar or tar in both meanings, but it always takes éis in dative),

With dative

Most prepositions expressing static position take dative, some of them can also take accusative in dynamic meaning (denoting target place of movement) – see #Stationary dative vs accusative of motion above.

  • aH ‘out of, from’
  • ag ‘at’
  • ar(L?) ‘on’
  • doL ‘to, for’
  • doL ‘from’
  • faL ‘under, beneath’ (not clear whether it takes accusative or dative in dynamic meaning)[2]
  • goN ‘with’
  • iN ‘in’
  • íarN, arN ‘after’
  • óL ‘from’
  • ósL ‘above’ (takes dative even in dynamic meaning)

With accusative

These typically take accusative in the singular, but dative in plural (except for mar, dar, gan which take acc. pl. too):

  • ar(L?) ‘onto’
  • iN ‘into’
  • idir, eidir ‘between’
  • (faL ‘under’?[2])
  • ganL ‘without’ – takes accusative also in plural
  • goH ‘to’
  • leH, léH ‘with’
  • reH, réH ‘towards, against’
  • reN ‘before’[3]
  • tar ‘across’
  • treL ‘through’
  • umL, bhaL, maL (faL?[2]) ‘about’

With genitive

Those prepositions are actually compounds of a preposition and a noun. The complement of those compound prepositions is put in genitive (because technically it is attributing another noun – the one being the second part of the preposition), and it will be either lenited or eclipsed – depending whether the second part of the preposition itself is in dative or accusative. For example i n-aghaidh meaning ‘against’ eclipse the following noun, because it means literally ‘in(to) the face of…’ and aghaidh ‘face’ is in accusative, eg. i n-aghaidh bhfear ‘against men’ and i n-aghaidh bhfir ‘against a man’, but in the meaning ‘in front of’ (as a stationary location) it lenites: i n-aghaidh fhir ‘in front of a man’ (eg. sitting at a table).

Personal forms

Forms with possessive pronouns

Forms with the article

  • aH ‘out of’: as an, as na + dative
  • ag ‘at’: ag an, gan, gun, agan, agon + dative (note the difference between gan fhior ‘at the man’, gan fhear ‘without a man’, gan tshúil ‘at the eye’ and gan shúil ‘without an eye’)[4]
  • ar ‘on’: ar an, ar na + dative or accusative
  • do ‘to’: don, dona + dative
  • do ‘from’: don, dona + dative
  • fa ‘under, beneath’: fan + dative (or accusative?[2])
  • goH ‘until, to’: gus an + accusative (gus an bhfear), anomalously just gus (gus bhfear), gus na + dative plural
  • goN ‘with’: gus an + dative
  • iN ‘in’: sg. san (sa), is in + dative or accusative, pl. sna, is na + dative
  • do-chum ‘to’ – generally not accepted in dán díreach, poets used go, gus an instead – even though it is an old compound preposition common in speech
  • leH ‘with’: leis an, leis na (note that it’s always leis before the article, even though it can be lais in conjugated 3rd. sg. form and in the meaning ‘also’)
  • reH ‘towards, against’: ris an + accusative, ris na + dative plural
  • reN ‘before’: rés an, rías an + accusative, rés na, rías na + dative plural
  • umL, bhaL, maL (faL) ‘about’: uman, bhan, man (fan)

Substantive verb: a-tú

The table below lists forms of the substantive verb ‘be’ (imperative), a-tú ‘I am’ (1st sg.) as they seem to have been accepted in the bardic standard.[5]

I assume that all non-impersonal forms used with the preverb do- are lenited, as seems to be the case (but is it always true in the earliest poems?).

1. sg. 2. sg. 3. sg. 1. pl. 2. pl. 3. pl. impersonal relative
present indep. a-tú a-táoi a-tá a-tám, a-támáoid a-tátháoi a-táid a-táthar
present dependent -fuilim, -foilim -fuile, -foile -fuil, -foil -fuilmíd, -fuileam,
-foilmíd, -foileam
-fuiltí, -foiltí -fuilead, -fuilid,
-foilead, -foilid
-fuiltear, -foiltear
present habitual bím bídh, -bí, -bíonn bímíd, -bíom bíth(e?) bíd, -bíad bíthear, (-bíothar?) bhíos
past indep. do bhá, do bhádhas do bhádhais do bhí, do bháoi do bhámar, do bhámair do bhábhair do bhádar do bás
past dependent -rabha, -rabhadhas -rabhadhais -raibhe, -rabha -rabhamar -rabhabhair -rabhadar -rabhas
past habitual do bhínn do bhíthe, (do bhítheá?) do bhíodh, do bhíoth do bhímís do bhíodh sibh, (do bhíthí?) do bhídís (do bítheá?)
future bíad, -biu, -biú bía bíaidh, -bía beimíd, -bíam (beithi? beidh?) beid, -bíad (beithear?) bhías (bhés?)
conditional do bheinn, -beinn do bheitheá do bhíadh, do bhíath do bheimís, do bheamáois do bhíadh sibh do bheidís (do beithe?)
present subj. bear, rabhar beis (béis?), rabhais , (beith?), rabh, raibh beam, rabham beithi, rabhtháoi bead, (beid?), rabhad beithear, (rabhthar?) bheas
past subj. do bheinn, do rabhainn do bheitheá, (do bheithe?) do bheith, do bheath,
do bheadh, do rabhadh
do bheimís, do bheamáois,
do rabhmáois
do bheith sibh do bheidís, do rabhdáois (do beithe?, do rabhtha?)
imperative bíor, bím bíodh, bíoth bíom bídh, bíthe bíod (bítear?)

The present tense dependent forms in -fuil, -foil are generally used after dependent particles: an bhfuil…? ‘is…?’, nach fuil…? ‘is … not …?’, go bhfuil… ‘that … is …’, etc. But in some contexts forms -tú, -táoi, -tá, -tám ⁊ -támáoid, -tátháoi, -táid-tád, are often used:

  • often after mar, mur: mar tá… ‘as there is…’,
  • after : má tá san tshíodh an tshleagh ghorm…[6] ‘If the blue spear … be in the mound’,
  • after preposition + relative aN: an t-ionadh a ttú ‘the place in which I am’, an t-athair ó ttú ‘the father from whom I am’, an t-adhbhar fa (a) ttú, an tí fa ttám ‘the matter about which I am (talking), the one about whom we’re (talking)’,
  • after N ‘where’: gá ttú, gá dú, gá ttám, gá dám (ris) ‘where am I, are we (with it)?’, often ‘in short, finally, no need to elaborate’ (summarizing speech) – in this phrase the -ttú, -dú, -ttám, -dám, etc. forms can alliterate either with t or with d (and are often spelt with d).[7]



The analytical forms (ie. 3rd sg. forms used with nouns and pronouns) of the copula are as follows[8]:

Tense / Mood Affirmative Interrogative Negative Negative interrogative Relative Negative relative other forms
present is (as) an H, nocha (nochan + vowel) nach asL (isL) nach go ‘that, so that’ + cop.: gur(b), gurab, gonadh(?)
‘although’ + cop.: giodh, gér(b)
‘if’ + cop.: másL
muna ‘if not’ + cop.: muna(b)
past do baL, robaL (dob’L, rob’L + vowel),
do badh, do budh, faH
narL, nárL (narbhL, nárbhL + vowel?)[9] níorL (níorbhL, níor bh’L + vowel), ní ba, ní badh,
nocharL (nocharbhL, nochar bh’L + vowel)
dob’L, faH, badhL, budhL nár (nárbh, nár bh’ + vowel) + cop.: gér(bh)L
(muna + cop.: munar(bh)L?)
future budh (badh) ní budh, níbaH bhus
conditional budh, do budh, robudh narL, nárL (narbhL, nárbhL + vowel?)[9] ní budh
present subjunctive rob nára (nárab + vowel) bhus go + cop.: gura(b), go mbaH
+ cop.: giodh
+ cop: madh
past subjunctive budh nába(dh) go + cop.: gomadh
+ cop.: gémadh
muna + cop.: munbadh

Note that in the past tense the shorter forms do ba, níor bh’, etc. cause lenition but the badh, ní badh forms don’t (but they still lenite in relative clauses). The present tense relative as also lenites.

The present form with gogonadh – continues Old Irish conid, but gur(a)(b) is more common. The latter form continues Old/Middle Irish corop which first was a variant of pres. subj. form, later reinterpreted as future, and later yet as present.[10]

Since the copula was always unstressed, the vowels were variously spelt. The present indicative historically was is but was often written as as, on the other hand the relative form (historically as) could be written as is, both were pronounced /əs/ by the classical times. Similarly the past and future forms were variously written as budh or badh.

Sealbhadh (synthetic conjugated forms)

The word sealbhadh is typically used for infix pronouns in classical bardic tracts, but it is also applied to the synthetic forms of the copula – it seems they were regarded as containing the infix pronouns (even if actually continued the old synthetic forms). The 3rd sg. masc. form aN is the infix pronoun (the copula didn’t distinguish gender originally, is was the OIr. 3rd sg. form).

The forms used in dán díreach are as follows:[11]

Singular Plural
1st am arN
2nd ad (at + vowel) abharN
3rd aN (as) (masc.)
as (fem.)
ad (+ consonant or vowel)

Those can be also prefixed with the 3rd sg. form is, as, eg. asam, isam for ‘I am’, asat for ‘you are’, etc. Sometimes as was considered the sealbhadh of 3rd sg. masc. copula too.[12]

If an adjective is the predicate, all the forms take singular adjective, except for 3rd pl. ad ‘they are’ which takes the plural form.

Some examples, note that the emphatic suffix -sa, -se, ⁊c. is attached to the predicate, and the full unlenited pronoun forms (sé, sí) are used after conjugated copula:

  • fear meallta Taidhg Í Uiginn mé = am fear meallta Taidhg Í Uiginn ‘I am a man deceiving/beguiling Tadhg Ó hU.’
  • ’s eisean meise = am seiseansa ‘I am him’
  • nar bean meise = narbam beansa (? ‘am I a woman?’ or ‘that I am not a woman?’)
  • gérb fear mé = géram fear ‘though I am a man’
  • Áodh tú = at Áodh ‘you are Aodh’
  • gérb é thú = gérad ‘though you are him’
  • Dé Domhnaigh eisean = a nDéisean Domhnaigh ‘it is Sunday’
  • gérb é é = géra ‘though he is him’
  • is usa na fir do mhealladh ioná… = as usa na fir do mhealladh ioná… ‘it is easier to deceive the men than…’ (as understood as the 3rd sg. form, standing for the whole na fir do mhealladh phrase)
  • Mairt isi = as Mairtsi ‘it is Tuesday’
  • d’fhearaibh sinn = ar nd’fhearaibh ‘we are from folks’
  • Áodh íad = ad Áodh ‘they are Aodh’
  • is usa na fir do mhealladh ioná… = asad usa na fir do mhealladh ioná… ‘the men are easier to deceive than…’


Predicative adjectives

Definite subject, indefinite predicative

copula⟩ ⟨predicative⟩ ⟨subject⟩:

  • is rí an fear ‘the man is a king’
  • is rí maith an fear ‘the man is a good king’

with adjectives more common:

  • is maith an rí an fear ‘the man is a good king, the king that the man is, is good’
  • is maith na fir an slúagh or is maithe na fir íad an slúagh ‘the host is good men’ vs is maith an slúagh é na fir ‘the men are a good host’

Both subject and predicative definite

When the predicate is definite, it is separated from the copular verb with a pronoun agreeing with the predicate.[13][gr 8]

copulaé/í/íadpredicative⟩ ⟨subject⟩:

  • mo theanga, is é m’arm-sa í ‘my tongue, it is my weapon’ (ie. it is the poet’s means to defend himself and by which he can do harm) – note that the first pronoun agrees with the predicate: masculine é and masculine arm, while the subject is feminine í ‘she’ standing for the feminine mo theanga ‘my tongue’,
  • is é an rí Conn ‘Conn is the king’,
  • is íad fir an bhaile an slúagh táinig… ‘the host that came was the men of the town…’,
  • giodh é críoch gach oighre ég ‘although the end of every heir is death’ – the masculine subpredicate é agrees with masc. ég ‘death’, and not with the feminine subject críoch gach oighre ‘the end of every heir’.
  • is í an tiobra gainmheach geal / Inis Fó[dh?]la na bhfinntreabh ‘the clear sandy well is the island of Ireland of the fair families’ (tiobra ‘well’ masculine according to McManus).

If the predicate is a singular collective noun, representing multiple people, the pronoun may be plural íad:

  • is í (or íad) an chlíar an chlíar táinig… ‘the group of poets (or clergy) that came was the group of poets/clergy…’ (í or íad can be used, because plural pronoun can be used for collective nouns).

When the subject is understood as a pronoun agreeing in gender with the predicate, it is often omitted (just like subject pronoun is often not expressed with other verbs, and like it wasn’t in Old Irish):

  • más é tal mo Thighearna ‘if it be my Lord’s will’ – no subject is explicitly expressed, é is part of the predicate.

But it appears occasionally:

  • is é an seanadh é más fhíor ‘it is the old tradition if it be true’ – the first é is part of the predicative (é an seanadh ‘the old tradition’), the second é is the subject ‘it’, and the subject is not repeated for the second copula más fhíor ‘if it be true’.

Demonstratives: ag so, ag sin, ag súd

While (é) seo, (é) sin, etc. could be used just like definite nouns in the copular sentences, there is an alternative popular construction with copular meaning (ie. ‘this is…, that is…’) in the classical language, but the ‘demonstratives’ may also carry an adverbial meaning (‘here is…, there is…’, pointing to a place or a general circumstance). It is formed by ag + the demonstrative + the noun in either nominative or accusative (sic!) – and when accusative is used, then the réim connsaine applies (ie. it is lenited).

For example:

  • ag sin mhnáoi ngil or ag sin bean gheal ‘that is a bright woman’ or ‘there is a bright woman’,
  • ag so an teangaidh nach tearc labhra / beannaigh, a shearc m’anma, í ‘behold this tongue [of mine] so excessive of speech (lit. that is not meagre of speech), bless it, o love of my soul’,[14]
  • ag so dearbhadh air ó Mhac an Bhaird ‘here is confirmation of it from Mac an Bhaird’,
  • Is í an bhreaghdhacht, gi bé i mbeath, / gaisgeadh is úaill is eineach; / ag sin tríar nár thrí lochta / do bhí i mBrían na breaghdhochta. ‘This Breagha-quality means in its possessor gallantry, pride, and hospitality; these were the three things – no defects surely! (lit. ‘these are [the] three things that were not three faults’) – to be seen in Brian of the Breagha-nature’.[15]

This is the construction which later gave rise to modern Irish and Scottish Gaelic copula-less forms like seo madra and seo cù ‘this is a dog’ or Munster and Connacht seo é an fear and Scottish seo am fear ‘this is the man’.

Nominative is much more common than accusative (only a few metrically fixed examples of accusative in bardic poetry) in this construction – but the accusative is older. It is derived from hypothetical Old Irish *aicc síu, *aicc sin ‘see here, see there’ with the (unattested) imperative *aicc ‘see’ of the verb ad·cí, the direct ancestor of classical faic ‘see’ – the predicate expressed after it being originally the direct object of the verb, hence the accusative and réim connsaine.[16]

In a similar manner – with either nominative or accusative – were used the phrases mo-chean ‘hail, fortunate, happy (is)’, mairg ‘woe to’, and mo-ghénar ‘happy, fortunate (is)’:

  • with accusative: Gá fhaisgin an fheadh do bháoi, / mo-chean mhnáoiLucky the woman who was seeing him as long as she was’,
  • with nominative: Mairg file ‘woe to the poet’.

Relative clauses

Rules of delenition

Classical Gaelic kept the Old Irish rules of delenition – when two homorganic (ie. produced in the same part of mouth) consonants came together, at least some of them resisted lenition even if that was required by the grammar.

The rules were different word-internally (typically in compound words where the second element would normally be lenited) and across word boundaries (eg. for adjectives following nouns in dative or a feminine nominative).

This section is based almost entirely on The Art of Bardic Poetry[gr 3].

Word internal


  • bh + b(h)b
  • b(h) + m(h)m (cráobh + mín = cráoimín; ibhmid ‘we drink’ read as if \*imid)
  • b + m(h)bm (cab + mín = caibmín)
  • b(h) + pp (garbh + peanngairpeann, )
  • m(h) + b(h)m (cam + bélcaimél – note difference with modern surname Caimbéal)
  • m(h) + m(h)m (cáomh + míncáoimín)
  • m + p(h)mp (gorm + portgormport, lom + peannloimpeann)

That is: sequences bhmh and mhbh change to m, but bmh to bm.

  • fh is not delenited by m or b (froimfhearthuin rhymes with poinnfhearfaidh; dob fhearr, ghobfhuar consonants with feadán and eagán)
  • f + m(h)/b(h)/p(h) → probably fm/fb/fp (but Mac Cárthaigh has no good examples; Ó hEódhasa in his RHG seems to claim in word boundary this can happen and delenition takes place)
  • p + b(h)p (ceap + beagceipeag)
  • when lenited, sometimes written f (glas + peannglaispheann and lus + portlusfort)

Dentals (coronals)

  • d(h) + d(h)d (bileadh + dreambileidream, fichead + dubhficheadubh; Díarmaddha rhymes with Díarmada),
  • d(h) + sts (an fhleadh-sa rhymes with an leath-sa and an cead-sa as if all ending in -eatsa)
  • d(h) + t(h)t (bhfleadhtunna rhymes with eaturra)
  • l(l) + dll (after short/middle vowel, gal or Gall + dlúithgallúith); → l (after long, gáol + dlúithgáolúith)
  • l(l) + -dha → either lda or la (after long vowel); either middle vowel + l(l)da or short vowel + lla (after non-long vowel; , sáoghalda being pronounced either sáoghalda or sáoghalla, but it affects how the vowel rhymes)
  • l + thlt – note this is in direct opposition to Munster -lth-
  • n(n) + dnn (after short/middle, ceann + deargceinnearg); → n (after long, cían + deargcíainearg)
  • n(n) + -dha – either nda or na (with the same vowel length shenanigans as lda, connda being either middle-length connda or short-length conna)
  • n + thnt
  • t(h) + d(h)t (slat + deargslaitearg, ráth + dlúith + tioghrátlúitiogh)
  • th + sts (leath-sa rhyming with cead-sa, as if leat-sa; also cloch and sgoth rhyme but an chloch-sa and an sgoth-sa do not because the latter is pronounced as if sgot-sa)
  • s + d(h)sd
  • s + t(h)sd (cnúas + tioghcnúaisdiogh, cneas + táobhcneasdáobh)


  • gh + g(h)g
  • g(h) + c(h)c (beag + ceannbeiceann)
  • c + cc (leac + crúaidhleacrúaidh)
  • c + gc (beag + ceannbeiceann)
  • broad ch + cchc (droch + cíalldroichcíall, cloch + ceanncloichceann)
  • broad ch + gchg (cloch + gércloichgér)
  • slender ch + c or gc (braich + glanbraclan)

Between words


b(h) delenites, and is delenited by, b(h), m(h), p

  • -m + bh--m b- (um beól ‘in my mouth’, um stands for i + mo here)
  • -mh + mh--mh m- (lámh mhín pronounced as if lámh mín)
  • -m + p(h)--m p- (um péin, um páis, um stands for i + mo here)
  • fh is not delenited by m or b (do fháisg fear láimh um fhear n-óir ‘tightened his grip on a golden (chess-)man’)
  • f + m(h)/b(h)/p(h) → probably fm/fb/fp (but Mac Cárthaigh has no good examples; Ó hEódhasa in his RHG seems to claim in word boundary this can happen and delenition takes place)

Dentals (coronals)

  • -d + th--d t- (… dod thógbháil féin ‘… to raise yourself’ with thógbháil pronounced tógbháil – this is a continuation of example in s)
  • -dh + th--dh t- (no examples)
  • -dh + dh--dh d- (no examples)

That is: d(h) delenites following d(h), t(h) and is delenited by d(h), t(h), n, l, s.

  • -l + th--l t- (feóil the pronounced as if feóil te)
  • -l + dh--l d- (súil dhearg pronounced as if súil dearg)
  • -n + dh--n d- (bean dhubh pronounced as if bean dubh)
  • -n + th--n t- (bean throm pronounced as if bean trom)

N does not delenite s, so -n + sh--n sh- (bean sháor and not \*bean tsháor, and also generally not \*bean sáor)

For t probably (but no examples in the tract):

  • -t + th--t th-
  • -th + th--th t-
  • -t + dh--t d-
  • -th + dh--th d-

(generally th delenites following dh, th and is delenited after d(h), t(h), n, l, s)

  • -s th--s t- (as tú as dóthcha… ‘it is you who are most likely…’ – after copula and dóthcha after relative copula would be expected to be lenited)
  • -s dh--s d- (see as dóthcha)

The examples for -d t-, -s t-, -s d- are from a single sentence: as tū a[s] dōthcha dod thógbháil féin ‘it is you who are most likely to raise yourself’.


  • -gh + g(h)--gh g- (Ó Dalaigh gil pronounced even if written Ó Dalaigh ghil)
  • -g + g(h)--g g-
  • -gh + c(h)--gh c-
  • -g + c(h)--g c-
  • -c(h) + ch--ch c- (glac crúaidh pronounced with unlenited crúaidh even if written glac chrúaidh)
  • -c(h) + gh--ch g- (cloch gér pronounced with unlenited gér even if written cloch ghér)


  1. Although bhfáinne fhir seems to also be allowed, see BST 200.11–30
  2. 2.0 2.1 2.2 2.3 See ABP[gr 3], pp. 238–239: 695 note and 713 note. IGT i lists fáoi (that is, 3rd sg. masc. form of fa ‘under, beneath’) among prepositions taking acc. with the dynamic meaning (the others being ann and air, ie. i ‘in(to)’ and ar ‘on(to)’; also note that 3rd sg. of dynamic i ‘into’ with acc. is inn, not ann), and does not list fa among variants of um, but it’s not clear whether the use of fa in the meaning ‘about’ is proscribed – it seems it is allowed in BST[gr 2] and the only example with verb of motion and accusative there (cuirfead m’fhalluing fa fhear) can be understood as ‘around’, variant of um, bha, ma (‘I will put my cloak around/about a man’ and not ‘beneath a man’). Mac Cárthaigh has no examples of fa + acc. meaning ‘(going) under’ in poetry.
  3. note that reN took dative in Old Irish, it switched to accusative during Middle Irish times; also note that reN causes eclipsis, eg. re bhfear ‘before a man’, while Modern Irish roimh and Scottish Gaelic ro both cause lenition
  4. See BST[gr 2] 213.31, p. 22.
  5. The list is based mostly on IGT III[gr 4] – the tract on irregular verbs, but also on Lé (the tables in the grammar sections as well as forms listed by McKenna in Aithdioghluim Dána[ed 1] in the vocab section), later forms listed in eDIL, conjugation of the substantive verb from NLI MS G 3, f. 73R2-73v15⁹⁹ (listed in the article Aspects of bardic poetry in the thirteenth century[gr 6]), and also consulted with the table in Trí Bior-ghaoithe an Bháis[ed 2]
  6. read: an tshleagh gorm because of delenition
  7. While after preposition + aN only alliteration with t is allowed, see BST[gr 2] 211 25 (p. 20, and note on pp. 148–149), IGT iii (Verbs)[gr 4] suggests -ttú in regular prepositional relative can also alliterate with d, but version from ms. C forbids it (An tionadh a dtu tinne amain duaim ris, ie. “An t-ionadh a ttú: tinne amháin (cóir) d’úaim ris” ‘An t-ionadh a ttú: only T is correct to alliterate with it’)
  8. See SnaG[gr 7] p. 417, also ABP[gr 3] p. 336, various places in BST[gr 2]
  9. 9.0 9.1 The only (prose, not dán díreach) example I found is dá marbhtha sagart ⁊ tú a n-ainbhfios ’nar shagurt é ‘if you were to kill a priest unaware whether he was a priest’ in Scáthán Shacramuinte na hAithridhe lines 2873–2874, p. 88 (also cited in lé glossary as “’nar<an ar, 2874”) but the tracts are pretty clear that the particle an changes to nar, nár with verbs requiring gur instead of go, cf. IGT i, ABP[gr 3] 275–278, §16 and BST[gr 2] 211.13–14 (both with the example nar mharbhus tú? ‘did I kill you?’; Mac Cárthaigh, p. 205, also notes Máire munbadh í Eamhear / nár geinedh no an ngeinfeadhear / ochta mar úan na mara / dá snúadh corcra as cosmhala ‘Máire – if she was not Eimhear – has there been born or will there be – (breasts like the foam of the sea) – two purple complexions more alike?’ in dán díreach) so this likely also is true for the copula.
  10. Cecile O’Rahilly (1966), Gurab, present indicative of the copula in Celtica vol. 7, pp. 33–37.
  11. As listed in BST[gr 2], Appendix I: Sealbhadh, pp. 252–253.
  12. Interpreted as as + aN with missing eclipsis?
  13. This usage existed already in Old Irish (where the subject of the copula was never expressed with a pronoun), for example in Old Irish ‘he is a man’ would be expressed as is fer (with the subject ‘he’ being included in the 3rd. sg. copula is ‘he/she/it is’), and ‘he is the man’ would be is hé in fer – note that ‘he’ here isn’t the subject, but rather a sub-predicate, part of the whole phrase hé in fer ‘the man’ (in Middle Irish the sentence could be is hé in fer hé, just like in Modern Irish is é an fear é – with the first pronoun being a part of the predicative, the second – the subject). This is more clearly visible in an example from Würzburg glosses: Críst didiu, is sí in chathir ‘Christ, then, he is the city’ with the feminine pronoun ‘she’ agreeing with feminine in chathir ‘the city’, thus literally the sentence is to be understood ‘Christ, then, he-is her – the city’ with ‘he is’ being represented by is itself. In Middle Irish, when separate subject pronouns started appearing, this could have been also expressed as *Críst didiu, is sí in chathir with being the subject ‘he’ referring back to ‘Christ’. In earliest Old Irish this sub-predicate wasn’t required, so *Críst didiu, is in chathir hypothetically would also be possible with the same meaning.
  14. Verses and translation from McKenna’s Aithdioghluim Dána, poem 70, Déana mo theagosg, a Thríonóid.
  15. Quatrain and translation taken from McKenna’s edition of Tánag d’Fhanaid an einigh.
  16. See R. A. Breatnach (1976), An Gléas Teaspáinteach in Éigse, vol. XVI, 3.


Editions of bardic tracts and articles about Classical Gaelic grammar

  1. Brian Ó Cuív (1973), The Linguistic Training of the Mediaeval Irish Poet, Dublin Institute For Advanced Studies
  2. 2.0 2.1 2.2 2.3 2.4 2.5 2.6 Lambert McKenna (1944), Bardic Syntactical Tracts, Dublin Institute For Advanced Studies
  3. 3.0 3.1 3.2 3.3 3.4 3.5 3.6 Eoin Mac Cárthaigh (2014), The Art of Bardic Poetry: A New Edition of Irish Grammatical Tracts I, Dublin Institute For Advanced Studies
  4. 4.0 4.1 4.2 Osborn Bergin (1916–1955), Irish Grammatical Tracts in Ériu, Supplement, vols. 8–17, Royal Irish Academy
  5. Eoin Mac Cárthaigh (2015), Sléagar agus ‘genitives lenited in special circumstances’ i bhfilíocht na scol in Aon don Éigse (Coimhín Breatnach, Meidhbhín Ní Úrdail ed.), pp. 239–245, Dublin Institute For Advanced Studies
  6. Pádraig Ó Macháin (2015), Aspects of bardic poetry in the thirteenth century in Aon don Éigse (Coimhín Breatnach, Meidhbhín Ní Úrdail ed.), pp. 91–125, Dublin Institute For Advanced Studies
  7. Damian McManus (1994), An Nua-Ghaeilge Chlasaiceach in Stair na Gaeilge: in ómós do P[h]ádraig Ó Fiannachta (Kim McCone ed.), pp. 335–445, Coláiste Phádraig, Maigh Nuad
  8. Damian McManus (2021), Identification copula clauses linking substantives of different gender in Early and Classical Irish in North American journal of Celtic studies, vol. 5, issue 2, pp. 214–241, The Ohio State University Press, doi:10.1353/cel.2021.0012

Editions of bardic poetry and other texts

  • Mícheál Hoyne (2018), Fuidheall Áir: Bardic poems on the Meic Dhiarmada of Magh Luirg, c. 1377 – c. 1637, Dublin Institute For Advanced Studies
  • Bardic Poetry Database – a searchable online collection of bardic poetry
  1. Lambert McKenna (1939), Aithdioghluim Dána: a miscellany of Irish bardic poetry, historical and religious, including the historical poems of the duanaire in the Yellow Book of Lecan, vol. 1 (introduction and text) and 2 (translation, notes, vocabulary), Dublin
  2. Geoffrey Keating (17th c.) Trí Bior-Ghaoithe An Bháis, edited by Robert Atkinson, published in 1890, Royal Irish Academy, Dublin