Fact-checked by Grok 2 weeks ago

Welsh syntax

Welsh syntax encompasses the principles and rules that govern the formation of sentences and phrases in the Welsh language, a Brythonic Celtic language primarily spoken in Wales and characterized by its Insular Celtic heritage. As of the year ending March 2025, there are an estimated 828,600 speakers of Welsh in Wales, representing about 26% of the population.^[1] Key features include a verb-subject-object (VSO) word order in finite declarative clauses, making it typologically distinct among Indo-European languages, as exemplified by sentences like Prynodd Elin dorth o fara ("Elin bought a loaf of bread"), where the verb prynodd ("bought") precedes the subject Elin and object dorth o fara.^[2] This VSO structure extends to head-initial phrasing across major categories, such as verb phrases and noun phrases, with restrictions on finite verb placement in embedded clauses and an elaborate inventory of clause-initial particles for functions like negation, questioning, and focus.^[3] A defining syntactic mechanism in Welsh is initial consonant mutation, where the initial consonant of a word changes (e.g., soft mutation, nasal mutation, or aspirate mutation) based on preceding syntactic triggers, such as the presence of a direct object or possessive, serving to mark grammatical relations without dedicated case morphology. Verb agreement is limited, typically occurring only with pronominal subjects in synthetic forms (e.g., Cerddon nhw adre "They walked home"), while lexical noun phrase subjects do not trigger agreement, reflecting a pro-drop tendency in colloquial varieties.^[2] Non-finite clauses, often introduced by particles like i or bod, play a central role in complex constructions, including infinitival complements and periphrastic tenses, contributing to the language's rich layering of auxiliary and serial verb structures.^[3] Dialectal and register-based variation further shapes Welsh syntax, distinguishing literary (or standard) Welsh from colloquial spoken forms, as in the colloquial Ddywedodd e ddim byd wrtha i ("He didn't say anything to me") versus the more formal Ni ddywedodd ddim wrthyf.^[2] These differences often involve particle usage, mutation patterns, and word order flexibility under emphasis, with ongoing research highlighting contact influences from English on contemporary syntax in bilingual communities.^[3] Overall, Welsh syntax exemplifies the analytic tendencies of modern Celtic languages while retaining archaic traits like VSO order, offering insights into historical change and typological universals.^[2]

Basic Clause Structure

Word Order

Welsh exhibits a canonical verb-subject-object (VSO) word order in affirmative declarative main clauses, a feature shared with other Insular Celtic languages but rare typologically, occurring in only about 10-12% of the world's languages. This structure is evident in both transitive and intransitive clauses without preverbal particles. For instance, a simple transitive clause is Gwelais i y llyfr ('I saw the book'), where gwelais is the inflected verb ('saw-1SG'), i is the subject pronoun ('I'), and y llyfr is the object ('the book'). An intransitive example is Cânodd y plentyn ('The child sang'), with cânodd as the verb ('sang-3SG') and y plentyn as the subject ('the child').^[4] Linguistic analyses supporting a surface VSO structure point to phenomena such as adverb placement and negation scope, which align with a flat or non-hierarchical clause structure where the verb precedes the subject directly. Adverbs, particularly manner adverbs, often appear immediately after the verb and before the subject in unmarked contexts, as in Gwelodd ddoe Emrys y ci ('Emrys saw the dog yesterday'), where ddoe ('yesterday') follows the verb gwelodd ('saw-3SG') but precedes the subject Emrys. Negation markers in simple clauses, such as ni, precede the verb but do not disrupt the postverbal subject-object sequence, as seen in Ni welodd Emrys y ci ('Emrys did not see the dog'), confirming the verb's initial position over the subject. These patterns argue against a derived VSO from deeper embedding, favoring a base-generated VSO order. Alternative generative approaches propose an underlying subject-verb-object (SVO) order, with the surface VSO derived via verb movement to a higher functional head, such as Tense or a split Complementizer domain, drawing parallels to Irish syntax where similar V-to-I movement occurs. In this view, the subject remains in situ within the verb phrase (VP), and the verb raises past it to satisfy agreement or licensing requirements, as analyzed in minimalist frameworks. Evidence includes coordination facts and remnant movement possibilities, where portions of the VP can be fronted, suggesting an initial SVO base before extraction. This movement-based account contrasts with flat VSO proposals by positing hierarchical structure akin to SVO languages.^[5]^[6] Overall, Welsh syntax is strongly right-branching, meaning heads precede their complements across major categories like verbs, nouns, and prepositions, leading to structures where modifiers and arguments follow the head, as in noun phrases like y ty mawr ('the big house'). This contrasts with left-branching tendencies in some analytic languages, where modifiers often precede heads more rigidly, such as in relative clauses, highlighting Welsh's synthetic, head-initial profile despite its VSO order.

Preverbal Particles

In Welsh syntax, preverbal particles such as mi and fe serve as affirmative markers in declarative clauses, appearing before finite verbs to indicate positive polarity and triggering soft mutation on the following element. The particle mi is typically associated with present and future tenses, while fe is used for habitual or past contexts, though both can co-occur with various inflected verb forms except the third-person singular present or imperfect of the copula bod.^[7] For example, the sentence "Mi aethon ni" illustrates mi with a past tense verb, translating to "We went," where aethon undergoes soft mutation from the root aeth. Similarly, "Fe welodd Siôn y ci" means "Siôn saw the dog," with fe preceding the mutated verb welodd (from gwelodd).^[7] For negation, the particles ni (pairing with mi) and nid (pairing with fe) introduce negative declarative clauses in formal or literary Welsh, also occupying preverbal positions and triggering soft mutation. These negators replace their affirmative counterparts to convey polarity, as in "Ni welais i y llyfr," meaning "I didn't see the book," where ni precedes the mutated welais (from gwelais).^[7] Likewise, "Nid oedd y ty mawr" translates to "The house was not big," with nid before the mutated copular form oedd (from bod). Syntactically, these particles are analyzed as occupying a functional head within the clause structure, such as the Tense Phrase (TP) or a Focus Phrase (FP), immediately preceding the finite verb in the verb-subject-object (VSO) order of declaratives. This positioning reinforces the finiteness of the clause and interacts with the base VSO word order by marking the left periphery without altering the core sequence.^[7] Both affirmative and negative particles consistently trigger soft mutation on the initial consonant of the following verb or noun, a hallmark of Welsh morphosyntax that signals their grammatical role. Historically, these particles developed from Middle Welsh pronouns and expletives, with mi deriving from first-person singular forms and fe from the third-person pronoun ef, evolving into dedicated complementizers amid the rise of verb-second (V2) effects around the 14th–16th centuries before a shift to stricter verb-initial order. This grammaticalization process, detailed in analyses of diachronic changes, involved phonological reduction and functional specialization, contributing to the loss of earlier V2 patterns.^[8] Dialectally, mi predominates in northern Welsh, while fe is more common in southern varieties, where particles are frequently omitted in colloquial speech, though mutation effects persist.^[7]

Negation and Questions

In Welsh, negation is primarily expressed through preverbal particles in formal or literary registers, where ni (before consonants) or nid (before vowels) precedes the verb and triggers soft mutation. For example, the sentence Nid yw hi'n dawel translates to "She is not quiet," with nid negating the copula bod and causing mutation on the following element. This structure aligns with the verb-subject-object (VSO) order of affirmative clauses, placing the particle immediately before the inflected verb. In subordinate clauses, na or nad serves a similar function, as in Credaf [na fydd hi yno] ("I believe [that she won't be there]").^[9] Historically, this preverbal negation follows Jespersen's Cycle, where an original marker like Middle Welsh ny(t) weakened and was reinforced by the postverbal element ddim, originally meaning "thing" or "anything."^[10] In colloquial or spoken Welsh, the preverbal particle is frequently omitted or reduced (e.g., to d or t), with negation relying more heavily on the postverbal adverb ddim, which functions as the primary clausal negator and often requires an n-word (negative polarity item) in subject, object, or adverbial position for full sentential negation. Structures like Dydy hi ddim yn dawel ("She isn't quiet") exemplify this, where ddim follows the weak negative form of the copula dydy (a contraction of does dy) and licenses n-words such as dim byd ("nothing") or byth ("never").^[11] Emphatic negation employs ddim postverbally to intensify denial, particularly with objects or adverbs, as in Methodd y ci y ci ddim ("The dog didn't see the dog at all"), where ddim underscores the negation beyond basic clausal marking. Non-finite negation uses peidio â (or softened beidio â) with infinitives, e.g., Peidio â gwneud dim ("Not doing anything").^[9] Yes/no questions in Welsh maintain the underlying VSO order but are marked by a preverbal interrogative particle or, in spoken varieties, rising intonation alone. In formal Welsh, the particle a introduces the question, as in A yw hi'n dod? ("Is she coming?"), where a precedes the inflected verb without altering the clause structure. Responses echo the verb or auxiliary, such as Ydw ("Yes, I am") or Nac ydw ("No, I am not"), reflecting polarity focus rather than simple affirmatives.^[12] Colloquial speech often drops a, relying on prosody for distinction, e.g., Yw hi'n dod? with upward intonation, though the particle y(d) may appear before vowel-initial verbs like Y mae yno rywbeth? ("Is there something there?"). Past tense questions use similar patterns, answered with do ("yes") or naddo ("no"). Wh-questions front the interrogative word or phrase to clause-initial position, followed by the VSO order, often with the particle a or yr in literary Welsh but frequently omitted in speech. Common wh-words include beth ("what"), pwy ("who"), pa ("which/what"), pryd ("when"), and lle ("where"), as in Beth wnaethoch chi? ("What did you do?"), where beth fronts and triggers soft mutation on the verb if applicable. Subject wh-questions avoid movement, maintaining VSO, e.g., Pwy welodd y ddraig? ("Who saw the dragon?"), while object or adjunct questions involve extraction, sometimes with resumptive pronouns in complex cases. Prepositional wh-questions front the entire PP, like I bwy welaist ti y ddraig? ("To whom did you see the dragon?").^[12] In embedded contexts, wh-questions use complementizers like y or mai, with pa often specifying alternatives, e.g., Gwn i [pa lyfr y buychoch] ("I know [which book you bought]"), or infinitival i-clauses for non-finite embeddings, such as Gwn i [i beth y gwnech] ("I know [what to do]"). Soft mutation may apply to the embedded verb, but wh-words like pa resist it.^[13] Modern spoken Welsh shows English contact influence in the formation of question tags, particularly through invariant forms borrowed into Welsh English varieties and occasionally calqued back into bilingual speech. Structures like nac yw e? ("isn't it?") or reduced tags using na(c) and oni(d) particles, e.g., Yn' bydd? ("won't she?"), parallel English invariant tags such as isn't it?, reflecting substrate transfer in tag questions for emphasis or confirmation. This convergence is evident in northern and southern dialects, where English-like tags attach to VSO clauses, as analyzed in contact linguistics studies.^[14]

Nominal Syntax

Noun Phrases

Noun phrases (NPs) in Welsh exhibit a right-branching structure, in which the head noun is followed by its post-nominal modifiers, such as adjectives and prepositional phrases, while a limited set of pre-nominal elements, including determiners, occupy positions to the left of the head.^[15] This organization aligns with the broader syntactic patterns of Insular Celtic languages, where pre-nominal functional categories like determiners and quantifiers precede the lexical noun, but attributive modification is predominantly post-head.^[16] The internal complexity of NPs is thus asymmetrical, with pre-nominal slots filled by a closed class of items that often interact with phonological processes such as initial consonant mutation.^[17] The definite article in Welsh takes three forms: y before nouns beginning with a consonant (e.g., y ci 'the dog'), yr before nouns beginning with a vowel or with h (e.g., yr ŷ 'the house', yr hen 'the old one'), and elided 'r following a vowel in rapid speech or certain syntactic contexts (e.g., mam 'r ŷ 'mother of the house').^[18] This article marks definiteness and typically triggers soft mutation on following feminine singular nouns, though masculine nouns remain unaffected (e.g., y gath from cath 'cat', but y ci unchanged).^[19] Indefinite NPs, by contrast, lack a dedicated indefinite article and are often bare (e.g., ci 'a dog'), though specificity can be conveyed by pre-nominal un 'one, a' or rhyw 'some, a certain' (e.g., un ci 'a dog', rhyw gath 'some cat').^[18] Possession within NPs is realized through two primary constructions: pre-nominal possessive pronouns, which agree in person and number and induce soft mutation on the following head noun (e.g., fy llyfr 'my book', where llyfr remains unchanged but would mutate in applicable cases like fy gath 'my cat' from cath), or post-nominal genitive apposition, where the bare possessor NP follows the head without a preposition (e.g., llyfr y plentyn 'the child's book').^[16] In the genitive construction, the possessor is typically definite and precedes any further modifiers of the head, maintaining the right-branching order.^[15] Demonstrative determiners are also pre-nominal and inflect for gender, number, and proximity: proximal forms include hwn (masculine singular, 'this'), hon (feminine singular, 'this'), and plural hyn; distal forms are hwnnw, honno, and hynny (e.g., y ci hwn 'this dog', often with the article for emphasis).^[20] These demonstratives function as definite determiners and may co-occur with the article in layered structures.^[17] Quantifiers occupy pre-nominal positions and frequently trigger soft mutation on the head noun. Universal pob 'every' and partitive rhai 'some' exemplify this, as in pob plentyn 'every child' (no mutation needed here, but applicable to mutable initials) or rhai cathod 'some cats', where plural agreement applies without mutation on the head.^[20] Other quantifiers like llawer 'many' or sawl 'several' follow similar patterns, integrating into the determiner layer and constraining the definiteness of the NP.^[15] Adjectives, when modifying the noun, uniformly follow the head in this right-branching framework (e.g., y ci mawr 'the big dog').^[16]

Adjectival and Predicative Complements

In Welsh, attributive adjectives typically follow the noun they modify, resulting in a postnominal order (NA). For example, tŷ mawr means "big house," where mawr ("big") appears after the head noun tŷ ("house").^[21] This placement is standard for the majority of adjectives, though a small class of prenominal adjectives, such as hen ("old"), gwir ("true"), and prif ("main"), precedes the noun and often triggers soft mutation on the following noun, as in hen gath ("old cat").^[22] When multiple postnominal adjectives modify a noun, they follow a fixed order prioritizing size, then shape or color, provenance, age, and finally quality, as in gardd fawr breifat ("large private garden").^[21] Attributive adjectives exhibit gender and number agreement through initial consonantal mutations, primarily soft mutation after feminine singular nouns. For instance, du ("black") becomes ddû following a feminine noun like gath ("cat"), yielding y gath ddû ("the black cat").^[23] In literary Welsh, some adjectives also inflect morphologically for gender and number, such as gwyn ("white") becoming gwen (feminine) or gwynion (plural), though this is largely obsolete in spoken varieties.^[23] Prenominal adjectives may similarly trigger soft mutation on the noun, depending on its gender and number, as in y lwyd wawr ("the grey dawn").^[22] Predicative adjectives, which ascribe properties to a subject, require the copula bod ("to be") and the predication marker yn, forming structures like Mae'r tŷ yn mawr ("The house is big").^[24] Here, yn precedes the adjective and triggers soft mutation, as seen in Mae Gwyn yn ddio ("Gwyn is lazy"), where dio mutates from diog.^[21] Unlike attributive uses, predicative adjectives do not agree in gender or number with the subject and remain uninflected beyond mutation.^[23] For emphasis or identificational focus, the adjective may front with yw or ydy, as in Mawr yw'r tŷ ("The house is big").^[21] Syntactically, predicative adjectives function as complements within a PredP (predication phrase) headed by yn, embedded under the copula bod in a small clause structure, as in mae Sioned yn hapus ("Sioned is happy"), where the AP (hapus) complements the Pred head.^[24] This configuration distinguishes predicative from attributive uses, with the former integrating into verbal copular clauses rather than nominal phrases.^[21] Comparative and superlative forms of adjectives are derived either synthetically with suffixes or analytically. Short adjectives often use -ach for the comparative (e.g., mawr "big" → mawrach "bigger") and -af or -aidd for relational or superlative derivations (e.g., mawr → mwyaf "biggest"), while longer adjectives employ mwy ("more") and mwyaf ("most"), as in mwy hapus ("happier") or y mwyaf hapus ("the happiest").^[23] These forms typically appear in predicative position with bod and yn, such as Mae Sioned yn hapusach ("Sioned is happier"), and may precede nouns in attributive contexts without mutation after mwy or mwyaf.^[21]

Verbal Syntax

Copular Constructions with 'Bod'

In Welsh syntax, the verb bod ('to be') functions primarily as an auxiliary and copular verb, linking subjects to predicative complements in equative, ascriptive, locative, and existential constructions. Unlike content verbs, bod exhibits defective paradigms, inflecting irregularly for tense, aspect, and limited person/number agreement, while enabling progressive and perfective formations central to verbal periphrasis.^[24]^[25] The present tense of bod displays a range of forms conditioned by person, number, and discourse context, with formal/literary variants differing from colloquial spoken ones. In the first person singular, wyf (or yw) appears in elevated registers, as in Rwyf i'n athro ('I am a teacher'), while colloquial speech favors dw i. The second person singular uses wyt, as in Wyt ti'n blentyn ('You are a child'), and the third person singular employs yw or mae (the latter for indefinite or non-pronominal subjects), yielding Mae Menai yn afon ('Menai is a river'). For third person plural, maen marks agreement with pronominal subjects, e.g., Maen nhw'n siarad Cymraeg ('They speak Welsh'), though mae often suffices in impersonal or broad contexts.^[24]^[25] Inflection on bod extends across tenses, with person and number marking more prominent in the present and imperfect than in future or preterite forms. The future tense uses the invariant stem bydd, as in Bydd hi'n athro ('She will be a teacher'), inflecting minimally for person in some dialects (e.g., byddaf for first singular). The preterite employs bu, also largely invariant, e.g., Bu e yn Lloegr ('He was in England'), while the imperfect draws from oedd, with fuller agreement like Oeddwn i'n blentyn ('I was a child'). These inflections support bod's role in tense marking, particularly in subordinate clauses where it periphrases non-finite verbs.^[24]^[25] As a copula, bod predicates identity or equation between a subject and a nominal, adjectival, or clausal complement, often in verb-initial VSO order characteristic of Welsh declaratives. In equative constructions, it equates definite NPs, as in Y meddyg ydy Gwyn ('The doctor is Gwyn'), where the copula ydy (a present variant) facilitates inversion for focus. Ascriptive uses link subjects to indefinite predicates, e.g., Mae Siôn yn athro ('Siôn is a teacher'), attributing a class or property. This copular function underscores bod's syntactic versatility, raising to C-position in main clauses to host preverbal particles.^[24]^[25] The progressive aspect relies on bod plus the particle yn followed by a verbal noun, forming periphrastic ongoing actions, as in Mae hi'n canu ('She is singing') or Roedd y plant yn llofa ('The children were playing'). This construction, obligatory for present progressives, integrates aspectual marking directly into the copular frame, with yn licensing the verbal noun as a predicative complement.^[24]^[25] Locative predication typically employs bod without the aspectual particle yn, directly taking a prepositional phrase (which may include the preposition yn meaning 'in') as complement to denote position or existence in place, e.g., Mae'r llyfr ar y bwrdd ('The book is on the table') or Mae Gwyn yn yr ardd ('Gwyn is in the garden'). Here, the copula predicates spatial relations, contrasting with progressive uses by omitting the aspectual particle.^[24]^[25] Existential and impersonal constructions further highlight bod's role via the third-person singular mae, introducing indefinite entities or events without a thematic subject, as in Mae'n bwrw glaw ('It is raining') or Mae 'na gath yn y tŷ ('There is a cat in the house'). This impersonal mae functions expletively, associating with weather expressions, existentials marked by 'na ('there'), or infinitival complements, filling a gap in finite verbal paradigms by enabling non-agreeing predication.^[24]^[25]

Non-Copular Verbal Constructions

In Welsh syntax, non-copular verbal constructions typically feature finite content verbs in a verb-subject-object (VSO) order, particularly for expressing past and future tenses without the copula bod. For instance, the sentence Gwelais i geffyl translates to "I saw a horse," where the finite verb gwelais (saw.1SG) precedes the subject pronoun i (I) and the mutated direct object geffyl (horse). Similarly, Mi welodd Rhiannon ddraig means "Rhiannon saw a dragon," illustrating the unmarked VSO pattern with the affirmative particle mi optionally preceding the verb. This structure holds in literary and formal registers, contrasting with more periphrastic forms in colloquial speech. Periphrastic constructions involving verbal nouns (VNs) are central to non-copular verbal syntax, often employing auxiliaries such as gwneud (do) combined with the VN to convey non-progressive aspects. An example is Gwnaeth Emrys agor y drws, meaning "Emrys opened the door," where gwnaeth (did.3SG.PST) functions as the auxiliary, followed by the VN agor (open). VNs in these contexts exhibit both nominal and verbal properties, serving as the core of the action while allowing complements like direct objects, as in Gwnaeth y ddynes agor y drws ("The woman opened the door").^[26] Transitivity in these constructions is realized without overt case marking on direct objects, which immediately follow the verb or auxiliary-VN complex and undergo soft mutation on their initial consonants. In Prynodd Dafydd y llyfr ("Dafydd bought the book"), the direct object y llyfr (the book) shows no mutation due to the definite article blocking it, but indefinite objects like beic in Prynodd y ddynes feic ("The woman bought a bike") mutate to feic. This mutation, triggered by adjacency to the preceding verb phrase, applies specifically to objects of finite transitive verbs and underscores the head-initial nature of Welsh phrases. Serial verb-like structures are rare in Welsh, with argument encoding favoring prepositional phrases over verb serialization for multi-argument verbs. For example, the verb siarad (speak) takes a prepositional object as in siarad â rhywun ("speak with someone"), avoiding a direct serial construction and instead integrating the recipient via the preposition â. This preference maintains the VSO integrity while accommodating complex arguments through adpositional means. Intransitive verbs preserve the VSO order without alteration, positioning the subject immediately after the verb. Consider Rhedon nhw i ffwrdd ("They ran away"), where the intransitive rhedon (ran.3PL) leads the pronominal subject nhw (they) and adverbial i ffwrdd (away). Another case is Cofiodd hi ("She remembered"), with the finite verb cofiodd (remembered.3SG) followed by the subject hi (she). These patterns ensure syntactic consistency across transitive and intransitive contexts in non-copular clauses.

Tense, Aspect, and Mood Formation

In Welsh syntax, tense, aspect, and mood are realized through a combination of synthetic inflections on finite verbs and periphrastic constructions involving the auxiliary verb bod ('to be') followed by a verbal noun (VN). Synthetic tenses are limited: the preterite is formed with endings such as -ais for first-person singular (e.g., gwelais 'I saw'), the future employs personal endings like -af (e.g., canaf 'I will sing'), and the conditional uses -wn (e.g., canwn 'I would sing'). These inflections apply primarily to the indicative mood and are more common in literary or formal registers, while the verb bod exhibits a fuller paradigm with six synthetic tenses.^[27] Periphrastic formations predominate in spoken and colloquial Welsh, especially for expressing future and conditional tenses, where an inflected form of bod precedes the VN, often linked by the particle yn (e.g., Byddaf i yn canu 'I will sing'; Byddwn i yn canu 'I would sing'). This analytic structure allows for nuanced tense marking beyond synthetic limits, with bod occupying the tense (T) head position and the VN serving as its complement in the aspectual or verbal projection. Aspects are exclusively periphrastic: the progressive uses yn + VN (e.g., Yr wyf yn darllen 'I am reading'), the perfective employs wedi + VN (e.g., Yr wyf wedi darllen 'I have read'), and habitual aspect is conveyed analytically via bod forms or particles like fe with the VN (e.g., Fe fyddaf yn canu bob dydd 'I sing every day'). A recent perfective variant involves newydd + VN (e.g., Yr wyf newydd ddysgu 'I have just learned').^[27]^[28]^[29] Moods, particularly the subjunctive, combine synthetic and periphrastic strategies; synthetic subjunctives draw from imperfect forms (e.g., canai 'that he/she sing'), while periphrastic subjunctives use i + VN in subordinate clauses (e.g., Eich bod yn dod 'That you come') or bod + yn + VN for potential readings (e.g., Byddaf yn caru 'I may love'). These constructions typically appear in embedded contexts like purpose or conditional clauses, with the auxiliary in T and the VN in a lower complement position. In modern Welsh, English contact has accelerated the shift toward analytic periphrastic forms over synthetic inflections, particularly in tense-aspect marking, as evidenced in varieties influenced by Welsh English where progressive and perfective uses align more closely with English patterns.^[30]^[31]

Complex Sentences

Coordination

In Welsh syntax, coordination links elements of equal status, such as noun phrases, verb phrases, or clauses, using a set of coordinators that include ac (and), neu (or), and ond (but). These coordinators typically introduce the second conjunct, preserving the underlying verb-subject-object (VSO) word order of Welsh. For instance, the sentence Prynodd Elin dorth o fara ac aeth adre illustrates clause coordination, where ac connects two independent VSO clauses meaning "Elin bought a loaf of bread and went home." Similarly, neu and ond function analogously, as in Mae Siôn yn canu neu yn dawnsio ("Siôn is singing or dancing") and Gwelodd Mair y ddraig ond ni welodd hi’r llew ("Mair saw the dragon but she did not see the lion"). Coordinated elements exhibit syntactic parallelism, requiring the conjuncts to match in category and structure to ensure grammaticality. Noun phrases, for example, coordinate straightforwardly with ac, as in Gwelais [i a Megan] geffyl ("Megan and I saw a horse"), where the coordinated subject triggers first-person singular agreement on the verb, reflecting left-conjunct agreement typical in pronominal cases. Verb phrases maintain this parallelism without gapping or deletion mechanisms common in English; instead, each conjunct retains its full VSO form, such as Mae Gwyn yn ddiog ac yn cysgu ("Gwyn is idle and sleeping"). In clause coordination, subjects may be omitted in the second conjunct if identical to the first, but the initial verb or auxiliary position is preserved, as in Deffrodd Mair a mynd allan ("Mair woke up and went out").^[2] This structure avoids the forward conjunction placement seen in some Indo-European languages, emphasizing the paratactic equality of conjuncts. Asyndetic coordination, lacking an overt coordinator, occurs primarily in lists of noun phrases, such as bwrdd, cadair a llyfr ("table, chair, and book"), where commas separate items before the final a or ac. Full asyndetic clause linking is rarer in modern Welsh but appears in historical texts, contrasting with the syndetic patterns dominant today. Correlative structures, involving paired elements like naill...neu ("either...or"), are uncommon; simple juxtaposition with a single coordinator prevails, as in Gwelodd Mair y ddraig ac fe welodd hi’r llew ("Mair saw the dragon and she saw the lion"), maintaining VSO integrity across conjuncts without complex embedding.

Subordination

In Welsh syntax, subordination is characterized by the absence of true subordinating conjunctions comparable to those in English or other Indo-European languages; instead, it relies heavily on preverbal particles and specific verb inflections to introduce dependent clauses. Particles such as i (used for purpose clauses, e.g., i'w weld ef 'to see him') and yn (marking aspectual or progressive interpretations in certain contexts) signal subordination without lexical content, often combining with verbal nouns to form non-finite structures. This system allows for embedding while preserving the language's core verb-initial properties, distinguishing it from coordination mechanisms that link independent clauses.^[32] Subordinate clauses in Welsh can be finite or non-finite, with the latter predominating in embedded contexts like complements or adverbials. Finite subordinates typically employ inflected verb forms that maintain verb-subject-object (VSO) order, similar to main clauses, but omit affirmative particles like mi or fe found in matrix declaratives (e.g., Credaf [y bydd Sioned yn dod] 'I believe [that Sioned will come]'). Non-finite subordinates, by contrast, use verbal nouns (e.g., darllen 'reading') preceded by particles such as i or the copula bod, resulting in subject-verb order (e.g., i Sioned darllen y llyfr 'for Sioned to read the book'). Tense in subordinates often aligns with the matrix clause, employing future or subjunctive forms for congruence, though non-finite constructions may lack explicit tense marking and rely on contextual interpretation.^[33]^[32] Historically, the development of subordination in Welsh traces back to Middle Welsh (c. 1150–1500), where adverbial clauses frequently lacked dedicated complementizers, evolving instead through the reanalysis of preverbal elements into particles. This shift contributed to the rise of a complementizer system, as detailed in analyses of verb-second loss, where pronominal subjects in specifier positions were reanalyzed as heads hosting particles like y or a in embedded contexts. By Early Modern Welsh, this resulted in a more standardized use of particles for subordination, reducing reliance on paratactic structures.^[34]^[35]

Relative Clauses

Relative clauses in Welsh are typically post-nominal modifiers that follow the head noun they restrict or describe, maintaining the verb-subject-object (VSO) word order characteristic of the language's main clauses. This right-branching structure allows for embedded clauses to provide additional information about nouns, such as in y ci a welais i ("the dog that I saw"), where the relative clause a welais i modifies y ci. Unlike many Indo-European languages, Welsh relative clauses do not employ a wide range of relative pronouns; instead, they rely primarily on invariant particles or zero marking.^[36] The primary relativizer in literary Welsh is the invariant particle a, which introduces restrictive relative clauses and triggers soft mutation on the following verb. For subject relatives, colloquial Welsh often omits a entirely, resulting in a zero relativizer followed by a gap where the subject would appear, as in y dyn welodd y ci ("the man who saw the dog"). In object relatives, a precedes a gap in synthetic verb constructions, exemplified by y ci a welodd y dyn ("the dog that the man saw"). For non-subject positions like prepositional objects or possessors, a resumptive strategy is common, involving inflected prepositions or pronouns, such as y fenyw werthodd Ieuan y ceffyl iddi ("the woman to whom Ieuan sold the horse").^[37] In periphrastic constructions, object resumption may involve agreement clitics, as in y car mae ’r lladron wedi ei ddwyn ("the car that the thieves have stolen"). Non-restrictive relative clauses, which provide supplementary information rather than defining the head noun, are set off by commas or intonation and typically use a in a manner similar to restrictive clauses.^[36] For adjunct positions within non-restrictive clauses, overt wh-words like lle ("where") or pryd ("when") are required, as in y blwyddyn honno, pryd gafodd Dylan Thomas ei eni ("that year, when Dylan Thomas was born"). Resumptive pronouns are more frequent in these constructions, particularly for non-subject gaps.^[36] Historically, Middle and archaic Welsh featured prenominal relative clauses, where the relative construction preceded the head noun, often using demonstratives like yr hwn ("who, which"), as in y dyn yr hwn a gafodd y wobr ("the man who got the prize"). This prenominal positioning is now obsolete in modern Welsh, having shifted to the current post-nominal pattern.^[36] Syntactically, the particle a functions as a complementizer heading a Complementizer Phrase (CP) in the relative clause, facilitating wh-movement for gap strategies while allowing resumption in island-sensitive contexts. This analysis aligns relative clauses with other wh-dependencies in Welsh, such as questions, where gaps predominate for subjects and direct objects but resumption occurs for obliques.^[37]

Focus and Topicalization

Fronting Constructions

In Welsh syntax, fronting constructions involve the displacement of a phrasal constituent (XP) to a preverbal position for discourse purposes such as topicalization or focus marking, contrasting with the unmarked verb-subject-object (VSO) order. These structures, often termed "abnormal sentences" in traditional grammar, allow any major constituent—including noun phrases, prepositional phrases, adverbials, or even verbs—to precede the finite verb, thereby highlighting the fronted element without altering the core VSO sequence that follows. This mechanism serves to establish a topic, convey contrast, or emphasize new information, and it is a hallmark of Welsh's flexible word order system.^[2] Topicalization typically features XP-fronting accompanied by a resumptive pronoun in the base position to maintain referential continuity, particularly when the fronted element is the subject or object. For instance, the sentence Y llyfr hwn, welais i ef ("This book, I saw it") fronts the object noun phrase y llyfr hwn ("this book") as the topic, followed by the verb welais ("saw") and a resumptive pronoun ef ("it") in object position. This construction is analyzed as left-dislocation, where the fronted XP adjoins to CP or occupies Spec-CP, with the resumptive pronoun ensuring grammatical relations; it is optional but common in narrative or explanatory contexts to separate the topic from the comment. Preverbal particles like a or dialectal variants (mi in southern Welsh, fe in northern) may co-occur, adding emphasis without triggering verb-second (V2) effects in modern Welsh.^[2]^[12] Contrastive focus, by contrast, often involves fronting without a resumptive pronoun, especially for adverbials, verbs, or adjuncts, to highlight opposition or corrective information. An example is Ddoe aethon ni ("Yesterday we went"), where the adverb ddoe ("yesterday") is fronted to contrast with an implied alternative time, followed directly by the verb aethon ("we went") and subject ni ("we"). Here, the fronted XP targets Spec-CP for focus projection, and the structure lacks resumption since the focused element is not coreferential within the clause; soft mutation on the verb (e.g., aeth from /gɛθ/) may occur under the XP Trigger Hypothesis, where preceding XPs induce mutation. This type of fronting is pragmatically driven, appearing in responses to questions or corrections, and differs from topicalization by its narrower scope on exhaustive or corrective interpretation.^[2]^[12] Syntactically, both topicalization and contrastive focus are motivated by A'-movement to Spec-CP, satisfying discourse requirements within a complementizer phrase that licenses the fronted constituent while preserving the verb's initial position in the matrix clause. In generative analyses, the fronted XP moves from its canonical post-verbal base position (e.g., VP or TP adjunct) to Spec-CP, with the finite verb raising to T(ense) or adjoining to C(omplementizer) in VSO languages like Welsh; this accounts for the consistent VS order post-fronting without full V2 inversion of the subject. Historical evidence from Middle Welsh reveals a more robust V2 pattern, where fronted XPs in Spec-CP triggered verb movement to C, as in examples like Riuedi mawr o sswydwyr a gyuodassant ("Great multitudes of warriors rose up"), reflecting a shift from V2 to strict VSO by the early modern period due to erosion of preverbal particles.^[2]^[38]^[8] Dialectal preferences influence fronting usage, with northern Welsh varieties showing a greater inclination toward fronting constructions, often reinforced by particles like fe or ddaru-clauses (e.g., Ddaru i weld y llyfr hwn "I did see this book"), which preserve archaic emphases and mutations more consistently than southern forms. This regional variation stems from historical retention of Middle Welsh features in northern speech, leading to higher frequencies of XP-fronting in colloquial northern registers for both topical and focal purposes, while southern Welsh may favor periphrastic alternatives or simpler VSO structures.^[2]^[12]

Cleft Sentences

Cleft sentences in Welsh are biclausal constructions used to express focus by highlighting a particular constituent, typically through a copula linking a fronted element to a relative clause. These structures allow for emphasis on subjects, objects, or adjuncts, distinguishing them from simpler fronting mechanisms by their exhaustive or contrastive interpretive effects. Unlike monoclausal fronting, clefts involve a specific syntactic frame that identifies the focused element as the unique or salient one satisfying the predicate in the relative clause. The prototypical it-cleft in Welsh follows the structure mae [focused constituent] y [relative clause], where mae is the present tense form of the copula bod 'to be', and y introduces the relative clause with a gap or resumptive pronoun corresponding to the focused element. For example, Mae'n Nhundee y byw hi translates to 'It is in Dundee that she lives', emphasizing the locative adjunct 'n Nhundee as the unique location of the event described in the relative clause y byw hi 'that she lives'. Another illustration is Beic brynodd y ddynes, meaning 'It was a bike that the woman bought', where the object beic 'bike' is fronted and the relative clause brynodd y ddynes 'that the woman bought' provides the background, with no soft mutation on the focused noun due to its extracted status.^[21] Syntactically, it-clefts are analyzed as involving movement of the focused constituent to a clause-initial position, akin to [Spec, CP] in generative terms, followed by the copula and a relative clause headed by y. This creates an identity predication where the initial element is equated with the trace or gap in the relative clause, triggering default third-person singular agreement on the verb regardless of the focused constituent's features, as in Fi welodd ddraig 'It was I that saw a dragon' (with welodd in third singular, not welais in first singular).^[39] This agreement pattern underscores the non-extractive nature of clefts, distinguishing them from wh-interrogatives or free relatives in unbounded dependency constructions.^[40] Wh-clefts, or pseudoclefts, in Welsh typically begin with a wh-phrase such as beth 'what', followed by the copula and an open clause, yielding structures like Beth a wnaethoch? Canu 'What did you do? Sing', where the wh-clause sets up the question and the following nominal provides the exhaustive answer. These constructions can reverse order for identificational purposes, as in Prifddinas Cymru yw Caerdydd 'The capital of Wales is Cardiff', emphasizing Caerdydd via the copula yw. Like it-clefts, they involve a relative-like clause but prioritize the wh-initial frame for highlighting the variable's value. Welsh clefts convey two main focus types: exhaustive focus, which identifies the fronted constituent as the unique satisfier of the relative clause (e.g., implying no other location in Mae'n Nhundee y byw hi), and contrastive focus, which sets the element against alternatives through intonation or context (e.g., rising tone on the copula in Caerdydd sy’n ddinas hardd 'It is Cardiff that is a beautiful city'). The exhaustive reading arises from the biclausal identity structure, while contrastive effects often align with prosodic cues.^[39] Historically, cleft constructions in Welsh developed during the Middle Welsh period (c. 1150–1500) as a means to express contrastive focus within emerging verb-second (V2) constraints, evolving from earlier patterns involving hanging topics and particles like a or y(d) that marked extractions. This rise paralleled the generalization of V2 orders, with clefts serving to front non-subjects while maintaining clause-initial verb placement in the relative portion.^[41] In modern Welsh English, a variety spoken by bilinguals in Wales, cleft sentences show increased frequency as calques from Welsh syntax, where native speakers transfer the biclausal focus strategy to English structures like 'It was in Dundee that she lives', reflecting substrate influence on emphatic constructions and contributing to syntactic variation in contact settings.^[42]

References

[1]
https://www.gov.wales/welsh-language-data-annual-population-survey-april-2024-march-2025-html
[2]
The Syntax of Welsh - Cambridge University Press & Assessment
1 - Introduction. pp 1-31 · 2 - Simple finite clauses. pp 32-67 · 3 - Non-finite verbs and infinitival clauses. pp 68-103 · 4 - Wh-constructions. pp 104-151 · 5 - ...Missing: key | Show results with:key
[3]
On the nature of Welsh VSO clauses | Request PDF - ResearchGate
Aug 6, 2025 · Kriang intransitive clauses have the structure Subject Verb. Transitive clauses have the structure Agent (Subject) Verb Patient (Object).
[4]
Principles and Parameters in a VSO Language: A Case Study in ...
This work provides an analysis of word order and clause structure in Welsh, within the context of a minimalist version of principles and parameters theory.
[5]
[PDF] Principles and Parameters in a VSO Language: A Case Study in ...
Roberts looks at additional empirical issues to argue against the verb rising to C for Welsh VSO, and in the process adopts Rizzi's (1997) split-C system to.
[6]
[PDF] The complementizer system in informal Welsh
First, preverbal particles occur before a finite verb. Second, there are differences in terms of clausal word order. Fe / mi occur only in normal-order clauses ...
[7]
(PDF) V2 to V1 in Welsh: The role of preverbal particles and fronted ...
Jan 21, 2025 · A peculiarity of V2 in Middle Welsh is the morphosyntactic marking of fronted constituents in positive declarative main clauses (PDMCs) ...<|control11|><|separator|>
[8]
(PDF) The syntax of Welsh negation - Academia.edu
Colloquial Welsh negation involves many complexities. Negative sentences require an n-word in subject or object position or as a post-subject adverb.Missing: scholarly | Show results with:scholarly
[9]
[PDF] A minimalist approach to Jespersen's Cycle in Welsh
Literary Welsh maintains preverbal ni(d) with optional postverbal ddim, and has therefore not participated in the change found in spoken Welsh. This ...Missing: scholarly | Show results with:scholarly
[10]
None
### Summary of Negation in Informal Welsh, Preverbal Particles, and Examples
[11]
https://users.aber.ac.uk/bmj/Ymchwil/Negative%20Preverbal%20Particles%20in%20Informal%20Welsh.pdf
[12]
(PDF) The syntax of Welsh “direct object mutation” revisited
Aug 6, 2025 · In Welsh (Tallerman 2006) , for example, the first consonant of a word W2 is 'mutated' if the preceding word W1 belongs to a phrasal constituent ...Missing: sources | Show results with:sources
[13]
[PDF] Heli Paulasto, Rob Penhallurick and Benjamin A. Jones Welsh English
3.4.1 Embedded inversion. Inverted word order in indirect questions, also known as embedded inversion. (EI), is a feature which is reported in numerous ...
[14]
Noun phrases (Chapter 5) - The Syntax of Welsh
Borsley , Maggie Tallerman and David Willis; Journal: The Syntax of Welsh. Published online: 1 October 2009. Configuring Mutation in Irish. Type: Article; Title ...
[15]
[PDF] 4 oun Phras Structur in lsh Louisa SadlER 4.1 Introduction Several ...
Possession is generally expressed in Welsh noun phrases by means of a bare d p np following the head noun and preceding any pp complements and modifiers: I will ...
[16]
[PDF] On the structure of {Welsh} noun phrases - HPSG Proceedings
Rouveret (1994) proposes a right branching structure, in which the possessor is a specifier in a nominal constituent from which the noun has been extracted ...
[17]
[PDF] The Article 1. There is no indefinite article ( a / an) in Welsh. e.g. cath
(ii). 'yr' is used in front of a vowel and in front of h. Remember that. 'w' and 'y' are vowels in Welsh. e.g. ysgol - yr ysgol - the school.
[18]
[PDF] SOME BASIC RULES OF WELSH GRAMMAR Cynnwys - Adject - BBC
* Notice that the singular form of the verb is used with all nouns - even with plural nouns. As 'Mi / Fe' are particles which denote the affirmative, they are ...
[19]
[PDF] Welsh Prenominals: at the Syntax-Morphology Interface - CSSP
The elements which come before the head noun are fairly few and include the definite determiner y and its variants yr, 'r,1 other determiners such as pob 'every ...
[20]
[PDF] The Syntax of Welsh
The authors cover a broad variety of topics, including finite and infinitival clauses, noun phrases, agreement and tense, word order, clause structure, dialect ...Missing: key | Show results with:key
[21]
[PDF] Welsh Prenominals: at the Syntax-Morphology Interface - CSSP
The standard position for attributive adjectives is postnominal, and the vast ma- jority of adjectives are limited to this position in non-literary Welsh. A ...<|control11|><|separator|>
[22]
[PDF] 13 Welsh - University of Oxford
https://doi.org/10.1017/9781108769617.020. Downloaded from https://www ... For further discussion of Welsh syntax, see Borsley, Tallerman and Willis (2007).
[23]
[PDF] Copular Clauses in Welsh - Aberystwyth University Users Site
The restoration of the copula shows that these Welsh examples have verb-second order, that is, they have fronted order, in which the copula occurs in medial ...
[24]
[PDF] The complexities of the {Welsh} copula - HPSG Proceedings
The Welsh copula has a complex set of forms reflecting agreement, tense, polarity, the distinction between main and complement clauses,.
[25]
The Welsh Verbal Noun - MDPI
The verbal noun in the modern, currently spoken p-Celtic language Welsh is of a different nature than any other word class known in Standard Average European ...
[26]
(PDF) Aspectuality in Welsh - ResearchGate
However, Modern Welsh differentiates synthetically between six tenses for the verb bod 'to be' and four for all of the others (save defective verbs) in the ...
[27]
[PDF] SUBJECTS OF VERBAL NOUNS IN EARLY MODERN WELSH
Embedded 'bod-clauses' thus have a. 18 The same is true for “embedded questions and clefts, embedded negative clauses and condi- tional and future clauses ...
[28]
[PDF] ASPECTUALITY IN WELSH - ojs tnkul
Aspectuality in Upper Sorbian is, therefore, predominantly expressed by different verb lexemes, tense variation or adverbs, and all Upper Sorbian gerunds can be ...
[29]
Mood in Welsh
### Summary of Mood Formation in Welsh (Subjunctive Focus)
[30]
(PDF) Welsh English Syntax: Contact and Variation - Academia.edu
Welsh English Syntax: Contact and Variation presents original research on the evolution and subsequent development of one of the Celtic-influenced varieties of ...
[31]
[PDF] On pseudo-non-finite clauses in Welsh - ZIL IPI PAN
Languages differ in how they employ finite and non-finite clauses. Welsh finite and non-finite clauses have a similar distribution.
[32]
Index
**Summary of Subordination in Welsh Syntax:**
[33]
Syntactic Change in Welsh: A Study of the Loss of the Verb-second
Author, David W. E. Willis ; Edition, illustrated, reprint ; Publisher, Clarendon Press, 1998 ; ISBN, 0198237596, 9780198237594 ; Length, 306 pages.
[34]
[PDF] Specifier–head reanalyses in the complementizer domain
Jul 23, 2007 · Willis, David W. E. 1998. Syntactic change in Welsh: A study of the loss of verb-second. Oxford: Oxford University Press. Page 32. 32. 1. My ...
[35]
Relativization strategies: NP accessibility in Welsh
Nov 28, 2008 · This paper is concerned with strategies for relative clause formation, both in the universal and language-specific sense.<|control11|><|separator|>
[36]
[PDF] The limits of resumption in Welsh wh-dependencies - David Willis
As would be expected, the syntax of relative clauses and other wh-constructions is very similar in. Welsh. However, there are some differences, one of which ...Missing: scholarly | Show results with:scholarly
[37]
[PDF] 9 Historical syntax - David Willis
In this chapter, we turn to the historical development of Welsh syntax, concentrating mostly on the Middle Welsh period (1150–1400) to the present-day.
[38]
[PDF] ON A FAMILY OF WELSH CONSTRUCTIONS
Three Welsh constructions are of interest here: wh-interrogatives, free relatives, and cleft sentences. Wh-interrogatives are quite like their English ...
[39]
On the nature of Welsh unbounded dependencies - ScienceDirect.com
Among Welsh UD constructions are wh-interrogatives, relative clauses, cleft sentences, free relatives and complements of adjectives like hawdd ⿿easy⿿. The ...Missing: scholarly | Show results with:scholarly
[40]
[PDF] Reconstructing the rise of V2 in Welsh - University of Cambridge
The pre-verbal particles a and y in Welsh positive declarative and relative clauses are exactly the kind of functional items that are stored as syntactic ...<|control11|><|separator|>
[41]
The rise of it-clefting in English: areal-typological and contact ...
Jul 1, 2009 · Many linguists have wanted to see a typological connection between rigidity of word order and clefting and sought to explain the spread of the ...