197
Erich Poppe DOI: 10.4312/linguistica.63.1-2.197-212
University of Marburg
poppe@mailer.uni-marburg.de
BETWEEN THE IMPRESSIONISTIC AND THE ARITHMETIC: 
THINKING ABOUT CRITERIA FOR THE STYLISTIC ANALYSIS 
OF EARLY MODERN WELSH PROSE*
1 RESEARCH QUESTION AND METHODOLOGY
This article is an exercise in empirical historical stylistics. It proposes to reflect on 
methodologies and criteria for a stylistic analysis of Early Modern Welsh prose, i.e., 
of works written in the Welsh language roughly between 1500 and 1700. Style is a 
concept difficult to define; it refers to phenomena of the linguistic organisation of a 
text’s surface, in the view of Biber/Conrad (2019: 16) specifically to the ‘distribution 
of linguistic characteristics [which are] frequent and pervasive in texts of the variety’ 
and to ‘features [which] are not directly functional; they are preferred because they are 
aesthetically valued’. For the historical period under scrutiny here, ‘Ciceronianism’ 
may be one such culturally dominant aesthetic preference. In the strict sense this means 
the imitation of Cicero’s periods, but in a looser sense a penchant for long and complex 
layered sentences (see, for example, Monfasani (1999), Robert (2011), Marsh (2013) – 
the literature on this topic is vast). This stylistic preference has already been noted for 
Welsh authors, for example by Davies (1995: 73) for Gruffydd Robert (c. 1527–1598), 
the writer in focus in this article.1 Proper Latin Ciceronian periods have a specific struc-
ture of cola and commata (Hofmann/Szantyr 1972: 732, Mueller 2007). The transfer 
of this concept to Early Modern ‘periods’ can be problematic since they may not nec-
essarily follow the same rules and be simply ‘long’, as pointed out, for example, by 
Lorian (1973: 159) for some sixteenth-century French writers and by Robinson (1998: 
105–119) in a spirited critique of much of sixteenth-century English prose: ‘The real 
English monster sentence is a sixteenth-century phenomenon, caused by the unsuccess-
ful grafting of Latin syntax on to English’ (Robinson 1998: 112). 
This article intends to provide some descriptive data for the reconstruction of the sty-
listic practice of an Early Modern Welsh writer, Gruffydd Robert, as a preliminary point 
* Research for this article was conducted within the Marburg research project ‘The Welsh Contribution to the 
Early Modern Cultures of Translation: Sixteenth-Century Strategies of Translating into Welsh’, led by Elena 
Parina and myself, a part of the German Research Council’s (DFG) Priority Programme 2130 ‘Cultures of 
Translation in the Early Modern Period’. Translations of the Welsh quotations are my own. Thanks are due to 
Raphael Sackmann and Oliver Currie for their detailed comments on earlier versions, to Paul Bryant-Quinn for 
helpful advice, and to the two anonymous readers for their suggestions for improvment. All remaining errors 
and infelicities are my own responsibility.
1 An anonymous reviewer kindly alerts me to the possibility that ‘Robert’ is probably a patronym and not a 
surname.
198
of reference for future larger-scale intertextual comparisons. It is based on the micro-sty-
listic analysis of his introduction to a catechism, which belongs to the larger text type of 
introductory paratexts. It has the methodological advantage of being manageable in size 
for a detailed exploratory investigation. The domain of micro-stylistics is the sentence. 
Sowinski (1999: 89–101) gives a helpful catalogue of linguistic features relevant for such 
micro-stylistic analyses: sentence length, sentence form (simple, complex, reduced/ellip-
tic; interruptions in sentence construction such as appositional and parenthetical phrases), 
order of constituents, sentence type (declarative, imperative, optative, interrogative), and 
variation of grammatical categories. Further sub-categories are the number of constitu-
ents in a sentence, the internal structure of constituents, and the number of coordinated 
and subordinated phrases, as well as the patterns of arrangement of main clauses and 
subordinate phrases relative to each other (see similarly Mehler (2005: 339–340) for the 
perspective of quantitative stylistics). Such criteria overlap with criteria for the measure-
ment of syntactic complexity (which is different from, for example, lexical complexity 
for which the choice and register of words would be considered). Based on Rescher’s 
(1998: 1) general definition of complexity – ‘Complexity is first and foremost a matter 
of the number and variety of an item’s constituent elements and of the elaborateness of 
their interrelational structure’ – Pallotti (2015: 118) specifies ‘structural complexity’ in 
linguistics as ‘a formal property of texts and linguistic systems having to do with the 
number of their elements and their relational patterns’. Thus, micro-stylistic analyses as 
well as measurements of syntactic complexity both refer to the number and arrangement 
of elements in a sentence and are therefore, at least in part, amenable to an arithmetic 
approach.2 On the other hand, empirical historical stylistics is a methodologically chal-
lenging field since it ‘perhaps inevitably combines the impressionistic and the arithmetic’ 
(Guillory 2017: 63) – the former derived from readers’ response to a close reading of texts 
and the latter from counting elements in texts and sentences (for a survey of the complexi-
ties of statistical stylistics or stylostatistics proper, see Tuldava 2005). Analysts therefore 
face the challenge in their presentations and interpretations of how to combine these two 
perspectives in order to be able to make meaningful statements about the style of a text.
2 GRUFFYDD ROBERT AND MORYS CLYNNOG
This micro-study is based on Gruffydd Robert’s introduction to Athravaeth Gristnogavl 
(‘Christian Doctrine’), Morys Clynnog’s adaptation of Diego de Ledesma’s Latin Doc-
trina Christiana (text: [Clynnog] 1568: [ii]-[vi], Lewis 1948: 4–6 with modernised 
spelling, capitalisation, and punctuation; background and sources: Bryant-Quinn 1998: 
7–19, 2000: 21). Robert prepared it for publication and printing in Milan in 1568 and 
contributed the paratext in the form of an address to Clynnog. In it, he highlights the 
importance of his work for the religious instruction of the Welsh, since such works had 
so far been lacking ([Clynnog] 1568: [ii]-[iv], Bryant-Quinn 2000: 25–27). Robert and 
his uncle Clynnog (c. 1520/21 – in or post 1581) spent most of their lives as Catholic 
exiles in Italy (Bryant-Quinn 2019, 2000, Williams/Bryant-Quinn 2019). In their view, 
2 For a discussion of syntactic complexity in Maurice Kyffin’s Deffynniad Ffydd Eglwys Loegr (1595), see Poppe 
(2022).
199
the Welsh language was the crucial medium for counter-reformation activities in Wales 
and for the education of its people, and this provides the background for the production 
of the Athravaeth (Price 2019: 188–189). 
The choice of Robert’s paratext for this study is motivated by two factors: first by its 
shortness which allows not only a comprehensive analysis, but also its comprehensive 
documentation, and second by the fact that Robert is considered in modern scholarship 
to be one of the Catholic writers of the time who, as highlighted by Bowen (1999: 22), 
were ‘interested in the influence of the Renaissance on language and style’ and whose 
works would therefore follow advanced contemporary stylistic conventions and ex-
pectations in their application to Early Modern Welsh prose.3 This is also reflected in 
his reputation as a Ciceronian, as a writer who was able to write in a Ciceronian style, 
which derives first of all from the fragment of his translation of Cicero’s dialogue Cato 
Maior de Senectute, transmitted in what is now extant of the sixth booklet of his Welsh 
grammar (probably printed some time after 1584) and intended as a stylistic inspira-
tion for contemporary Welsh prose authors. Davies (1995: 73) wrote of this incomplete 
translation (see also Griffith 1953–58: 20, 1966: 287):
Enough, however, survives to make clear the way in which Gruffydd Robert 
sought in his translation to capture the periodic style of Cicero’s Latin with its 
finely balanced correspondence of phrases and subordinate clauses.
Even more instructive in our context is his assessment of the style of a sentence 
he quotes from the introductory non-technical dialogue in the first booklet of Robert’s 
grammar, published in Milan in 1567, the year before the publication of the Athravaeth, 
which in his view ‘illustrate[s] Gruffydd Robert’s Ciceronian sense of style and peri-
odic cadence in his own Welsh writing’ (Davies 1995: 75).4
3 SOME ARITHMETIC: SENTENCE LENGTH AND SYNTACTIC DEPTH
The corpus of this study consists of the sentences of Robert’s introductory paratext 
([Clynnog] 1568: [ii]-[vi]). These are given in the appendix, sentence by sentence, each 
sentence numbered and accompanied by an English translation. The overall number of 
sentences in the paratext and the problems with their demarcation are discussed below, 
as are the details of the notation for their schematic presentation. 
A first impression a reader may take away from Robert’s paratext is probably one 
of ‘complexity’. Features which would contribute to this impression are the length 
of some sentences, some layered subordination, and repeated use of parallelism, i.e., 
of syntactically equivalent elements in two or more consecutive parts of sentences 
3 For Robert’s ideas about the necessary improvement of Welsh, see Griffith (1953-58) and Bryant-Quinn (2000). 
For the ideas of another Catholic author and translator, Robert Gwyn, who was not influenced in the same way 
by humanist and Renaissance ideals, see Bowen (1999: 28-42), Poppe (2019), and Parina/Poppe (2021).
4 Bowen (1999: 13) suggests that the translation of extracts from St John Chrysostom’s Homilies, in [Clynnog] 
(1568: 3-4), ‘are most likely Gruffydd Robert’s work’: these require separate linguistic analysis.
200
(compare Ostrowicz 2003). However, it needs to be acknowledged that any such sub-
jective reactions and the constitution of style in reception (compare Wesche 2015: 383) 
are shaped by the reader’s background and linguistic experiences and perhaps distorted 
by the historical distance between the early-modern text and the modern reader.
The arithmetic approach on the other hand would aim to establish a more objective, 
quantitative analytic framework. Sentence length, the number of words in a sentence, 
and syntactic depth, the number of subordinate phrases in a sentence, may be strong pa-
rameters in the context of sixteenth-century Ciceronianism. The former quickly comes 
up against a methodological challenge, namely of how to demarcate relevant sentences, 
or sentence-like units.5 ‘Sentence’ in the modern linguistic sense is probably not the ap-
propriate unit to capture Robert’s (and his contemporaries’) perception of the building 
blocks of texts. This has been forcefully argued for (most) Early Modern English prose 
by Robinson (1998) – see also Croll (1966: 231) – and also for Early Modern German 
prose on the evidence of its punctuation, for example by Stolt (1990). Evidence from 
Robert’s grammar of Welsh indicates that he thought in terms of traditional rhetorical 
units (for these, e.g., Rinas 2022: 118–121), rather than of grammatical ones. He in-
troduces the concepts of the rhessụm cyflaụn/perphaith (‘complete utterance’, corre-
sponding to oratio/sententia perfecta), whose beginning is defined by a capital letter 
and whose end by a punctus (Robert 1939: 18, 65). Its sub-units are marked by a colon 
and a comma respectively: ‘Gụahannod […] a ḍengys ressum megis hanner perphaith. 
Rhagụahānod syḍ […] yn arụyḍo bod yn y ḷe hụnnụ uahā, ond amherphaith’ (Robert 
1939: 65, ‘A colon […] marks an utterance as half-complete. A comma […] shows that 
in this place is a break, but [an] incomplete [one]’). His example proves that a complete 
rhessụm does not necessarily correspond to a modern sentence: this is a couplet which 
consists of an asyndetic sequence of three main clauses: ‘Ti yu’r gụan, taụ ar y guir: ar-
rian da a ụrandeụir.’ (Robert 1939: 65, ‘you are the weak one, speak not the truth: good 
money is heeded.’).6 The couplet represents the period, and Ti yu’r gụan, taụ ar y guir its 
first colon and arrian da a ụrandeụir its second; Ti yu’r gụan and taụ ar y guir respec-
tively are commata. In this poetic example, the length of the period and the presence or 
absence of subordinated phrases are not an issue. The rules on capitalisation and the use 
of the punctus suggested by Robert, and thus his implied understanding of the nature 
of syntactic units, have an important bearing on syntactic analysis and on its arithmetic 
presentation when the sentence is a relevant unit of analysis (see below). But it needs 
to be taken into account that these may have been only imperfectly implemented in his 
introduction to the Athravaeth – probably partly due to the Italian typesetters. Eight 
sentences are demarcated by an initial capital letter and closed by a punctus or a ques-
tion mark – these are sentences (1), (2), (3), (4), (6), (7), (9), and (12) in the schematic 
5 There is the further issue of what constitutes a word; for example, is iụ ‘to their’, the combination of the 
preposition i ‘to’ and the possessive pronoun 3rd plural (modern spelling i’w), to be analysed as one word or as 
two words? For the purpose of this paper, a word is provisionally defined as a typographical unit and iụ therefore 
as one word between spaces.
6 A variant of this couplet is attested in a poem by Iorwerth Fynglwyd (fl. 1485–1527) addressed to Rhys ap Siôn 
o Lyn-Nedd: ‘tydi’r gwan, taw di â’r gwir—/arian da a wrandewir’, see Jones & Rowlands (1975: 14).
201
presentation below and in the appendix. Sentence (5) begins with a capital letter, but is 
not closed by a punctus, i.e., the capital initial of (6) is not preceded by a punctus, so 
ignoring capitalisation, (6) could be read as a main clause coordinated with (5). Lewis 
(1948: 5) in his normalized edition of the text opts for a new main clause. There may be 
an internal punctus in (5) followed by a small letter – Lewis (1948: 5) inserts a comma 
in this position, and this interpretation is provisionally followed here as well. Sentence 
(8) begins with a small initial after a question mark and is closed by a punctus. Sen-
tences (10) and (11) are the most problematic ones. (10) begins with a small letter after 
a punctus and the conjunction canys, so ignoring the punctus it could also be read as a 
coordinated main clause belonging to sentence (9).7 There is an internal semi-colon in 
(10) which separates a subordinate clause from the preceding text to which it semanti-
cally belongs. Probably on semantic criteria, Lewis (1948: 6) inserts a comma instead of 
the semi-colon and places a full stop after the subordinate clause, and he begins not only 
a new sentence, (11), but also a new paragraph. The closing punctus at its end is in the 
text. Alternatively, (10) and (11) could be read a long rhetorical unit with two complex 
cola, perhaps even connected to sentence (9). Lewis (1948: 4–6) divides the text into 
12 sentences, and this is the internal structure provisionally accepted here. However, 
sentence (5) could be taken as two separate main clauses; sentence (6) could be joined 
to sentence (5) to form a rhetorical unit consisting of three coordinated main clauses (or 
of two, if (5) is divided up); sentences (10) and (11) could be read as one unit, perhaps 
even in conjunction with (9). This leaves modern readers with considerable uncertain-
ties about Robert’s intentions and also introduces fuzziness in the attempt to measure 
‘sentence-length’.
Robert’s paratext contains 575 words and can tentatively be divided into 12 (or 13 or 
fewer) ‘sentences’, depending on readers’ balance of semantic or syntactic criteria and 
their interpretation of punctuation.8 Based on a division of the text into 12 sentences, 
sentence-length varies from 24 to 91 words. If sentence (5) is separated into two main 
clauses, the shortest sentence will contain 13 words; if sentences (9), (10), and (11) are 
taken as one unit, they will contain altogether 103 words. The average number of words 
per sentence, based on 12 sentences, is about 48 words; the median is about 43 words. 
Both values hide text-internal variety. For the question of authors’ ‘Ciceronianism’, the 
attested maximum values of 91 in two sentences are perhaps more revealing. Here, the 
first two sentences stand out for their length (unless (9), (10), and (11) are accepted as 
a rhetorical unit with 103 words).
Prototypical Ciceronian sentences are not only long, but also layered. The inter-
nal structure of sentences in the paratext is another issue of interest, specifically their 
syntactic depth, i.e., the number and arrangement of subordinate finite and non-finite 
phrases below the level of a main clause. The following schematic presentation is 
7 The issue of the status of the causal conjunction canys as coordinating or subordinating needs further scrutiny; 
it is here provisionally taken as a coordinating main-clause conjunction, in accordance with its classification in 
grammars of Modern Welsh, compare Thomas (1996: 461, 466).
8 In a further step of refinement, account could be taken of the ratio of different sentence-types, i.e., declarative, 
interrogative, optative, exclamative.
202
intended to give an overview of syntactic depth in Robert’s paratext, based on its divi-
sion into twelve sentences. This explicit format is, however, practical only for short 
texts. The word-count for each sentence is given in parentheses. The main clause is 
marked ‘0’, numbers identify clauses and phrases on each syntactic level, syntactic sib-
lings are distinguished by subscript numbers. Robert’s sentence 11 is analysed in detail 
below. In order to explain the system of notation, I present first the first sentence in full 
and then the schematic presentation with explication:9
(1) 
[11]  VEdi ymy ḍarlain ych ḷyfr, o‘r athraụaeth Gristnogaụl, a chanfod ynḍo me-
gis egin pob pụnc hyles i gristion ụrtho, i gadụ‘r enaid, 
[2]  a ḍarfu i ḍuụ i ụneuthur ar i lun, ai ḍelụ: ag a rybrynnoḍ Crist ai ụerthfaụr 
ụaed: 
[0]  e laụenychoḍ fynghalon 
[12]  ụrth ụeled tryssor mor ụ[e]rthfaụr yn yr iaith gymraeg: 
[X]  a maint 
[X11] syḍ o eissie cyfrụiḍid ar phorḍ Grist, yn gyphredinol ymysc gụyr yn gụlad: 
[X]  a‘r plant yn crio am fara 
[Y]  (mal y mae‘r prophụyd yn ḷefain) 
[X]  heb fod neb, 
[X12] ai tyrr iḍynt ag ai rhyḍ heb i ụenụyno
[After I had read your book on Christian doctrine, and found in it as it were in a nut-
shell every point necessary in order for a Christian to preserve the soul which God 
has made in His image and His likeness, and which Christ bought with His precious 
blood, my heart rejoiced to see such a precious treasure in the Welsh language: con-
sidering how great is the general need for guidance in the way of Christ among the 
men of our country, and the children crying out for bread (as the prophet exclaims) 
while there is none who breaks it for them and gives it without poisoning it.]
(1) 11–2–0–12–(X–X11–X–Y–X–X12)   (91 words) 
 (X = parenthetical observation; Y = parenthetical source marker)
This summarizes the following information: The main clause 0 is preceded by a 
subordinate phrase 11, on which another subordinate phrase 2 depends; syntactic depth 
in the field preceding the main clause is 2. The main clause is followed by a subor-
dinate phrase 12 and by a complex parenthetical observation X into which a further 
syntactically unconnected parenthetical remark Y is inserted. Syntactic depth in the 
field following the main clause is strictly speaking 1, to which the parenthetical phrase 
adds another layer. The phrases at level 1 before and after the main clause constitute 
syntactic siblings; coordinated syntactic siblings of the same syntactic class are ignored 
9 All quotations from Gruffydd Robert’s text are reproduced diplomatically from the digital facsimile with all 
printing errors uncorrected and unmarked.
203
for the purpose of this presentation, but will impact on syntactic complexity and stylis-
tic effect (see below).
(2)   0–11–21–11–22–X–0–12–23–31–23–32  (91 words)
 (connective in sentence-initial position; X = parenthetical explanation)
(3) 0–2–1  (24 words)
(4) 0–1–2  (40 words)
(5) 01–1–02  (32 words)
 (alternatively, two coordinated main clauses: 13 + 19 words, or (5) + (6) = 
67 words)
(6) 0–1  (35 words)
(7) 0–1–21–22–3  (45 words)
(8) 0–11–21–22–23 –0–12–0  (53 words)
 (connective in sentence-initial position)
(9) X–0  (25 words)
 (X = left-dislocated)
 (alternatively, (9) + (10) = 57 words, (9) + (10) + (11) = 103 words)
(10) 0–1  (32 words)
 (alternatively, (10) + (11) = 78 words)
(11) 0–1  (46 words)
(12) 0–11–21–0–12–22–23  (61 words)
 (connective in sentence-initial position)
There is some variation in syntactic depth in this short text, with main clauses plus 
one subordinate phrase in (6), (10), and (11), and a main clause plus a left-dislocated 
phrase in (9), besides more layered sentences as in (1), (2), (7), (8), and (12). The maxi-
mal syntactic depth in the field preceding the main clause is 2 in (1); in (2), (8), and 
(12), subordinated phrases with a syntactic depth or 2 and 3 respectively are inserted 
between a sentence initial connective and the rest of the main clause. The maximal syn-
tactic depth in the field following the main clause is 3 in (7). Subordinate phrases frame 
the main clause in (1) and mutatis mutandis in (2), (8), and (12), the four sentences 
which also rank highest with regard to their word count.
A focus on syntactic depth, however, potentially hides other significant micro-sty-
listic features of individual sentences on the level of constituents, for example their 
number and the patterns of parallelism and coordination, partly reflected in the low 
value of syntactic depth in relation to the number of words, as will be seen in the next 
section in the discussion of sentence (11), with a syntactic depth of 1 and a length of 
46 words. There is the further complication that a schematic presentation cannot easily 
present syntactic ambiguities when more than one syntactic analysis appears possible, 
briefly mentioned above with regard to the status of canys as subordinating or coordi-
nating in (6) and (10).
204
4 SOME MICRO-STYLISTICS IN ACTION
Robert’s sentence (11) with its syntactic depth of 1, a conditional clause attached to a 
verbless main clause, looks deceptively simple:
(11) 
[0]  gụyn i byd trụy gymru, 
[1]  pe parent ymhob eglụys ụrth aros y gụasanaeth, ne ar osteg ypheren, gartref 
ymysc tylụyth y ty i difyrru‘r amser ag ymhob cyniḷeidfa i ḍiḍanu‘r bobl, 
ḍarlain hụnn ne‘r cyfryụ ymadrodion a gadel i phorḍ henchụedlau coegion, 
a choụydau gụenheuthus, celụḍog.
[It would be a great blessing throughout all Wales if they [the Welsh people] 
made a habit, in every church while waiting for the service to begin, or during low 
Mass, at home among the household to shorten the time, and in every assembly 
to comfort the people, of reading this book or similar material and have done 
with old, false legends and flattering, lying cywyddau. (Translation adapted from 
Bryant-Quinn 2000: 26)10]
A closer look reveals two micro-stylistic features which are concealed by simple 
measurements of words per sentence and syntactic depth. The first is the repetition of 
syntactically equivalent elements in two or more consecutive parts of sentences, akin to 
the rhetorical figure known as compar or parison, the use of similarly structured phrases 
or clauses (McDonald 2007: 39). Relevant instances in this sentence are ḍarlain … a 
gadel ‘reading and having done with’, henchụedlau coegion, a choụydau gụenheuthus, 
celụḍog ‘false old legends and flattering, lying cywyddau’, and the long sequence of 
adverbial expressions spanning ymhob eglụys … i ḍiḍanu‘r bobl with further internal 
parallelism of i difyrru‘r amser ‘to shorten the time’ and i ḍiḍanu‘r bobl ‘to comfort 
the people’. Another noteworthy feature is the separation of two syntactically closely 
related elements, the finite verb parent and its objects ḍarlain … a gadel, by the inter-
vening long adverbial sequence (underlined). 
A search for parison and separation in other sentences of the paratext reveals that 
these features are not restricted to sentence (11). Parallelism in some form occurs in 
all sentences and contributes to their overall length. In sentence (1), for example, two 
coordinated verbal-noun phrases are contained in phrase 11 and two coordinated rela-
tive clauses in both 21 and X1 – altogether amounting to 52 of the 91 words of the 
sentence. In (2), phrase 22 consists of a sequence of four coordinated indirect questions 
(22 words) involving contrast; in (7), phrase 4 consists of a sequence of six coordinated 
indirect questions (32 words); in (8), a sequence of three coordinated objects in the 
main clause express semantically related concepts: ai diogsụrth eisteḍach, ai bustlaụl 
sertheḍ, ag ai smala gyfeḍach ‘their tardy lolling about and their foul obscenity and 
their vain merriment’. Here, parallelism overlaps with synonymia, the repetition of 
10 Cywydd, pl. cywyddau, a Welsh metrical form which consists of rhyming couplets.
205
(near-)synonymous words (see Adamson 2007). Other examples occur, for example, 
in sentences (1), ar i lun, ai ḍelụ ‘in his own image and his likeness’, and (2), gasclu 
yn grynno, a dosparth yn drefnus, ag yn eglur ‘assembled compactly and arranged or-
derly and clearly’. A separation of syntactically closely related elements by sometimes 
extended adverbial phrases is found in altogether four sentences. In sentence (6), for 
example, the finite verb gaant and its object y pethau are separated by an adverbial 
phrase of 16 words (underlined):
(6) 
[0]  Canys yn ych ḷyfr chụi yma nhụy a gaant oi ḍysgu yn haụd, meụn ychydig o 
amser, a thrụy ychydig help, a ḷai o gost, y pethau 
[1]  syḍ angenrheidiol iụ gụybod, i hen ag ifanc.
[Since in this book of yours they will find, to teach them easily, in little time and 
with little help and less cost, the things that are necessary to know for old and 
young.]
This adverbial phrase could have been placed at the end of the sentence. In other 
cases, probably no alternative slots were easily available for the placement of the ad-
verbial phrases, but at the same time their length, resulting from the accumulation of 
parallel elements, is the author’s stylistic decision – resulting in the concomitant wide 
separation of syntactically closely related elements.
5 SOME CONCLUDING REFLECTIONS
In the two preceding sections, two approaches to the micro-stylistic analysis of Early 
Modern Welsh texts ‘between the arithmetic and the impressionistic’ were exempli-
fied, the quantitative analysis of sentence length and syntactic depth and the ‘impres-
sionistic’ qualitative analysis of individual sentences in search for distinctive stylistic 
features. Sentence length and syntactic depth varies in the paratext, and the resulting 
variety may be another feature that could be productively explored. For the question 
of Welsh Ciceronianism, the attested maximal value of 91 words in two sentences is 
probably more instructive than the overall variation: Robert produced long sentences, 
but did not do so consistently. Significantly perhaps, his first two sentences are long 
and fairly complex, as if he wanted to signal his ability to inscribe himself into a con-
temporary valued stylistic register. In a similar paratext, Roger Smyth’s introduction to 
Crynnodeb o aḍysc Cristnogaụl (1609), his adaptation of the catechisms of the Jesuit 
Petrus Canisius, the first two sentences are among the three longest ones.11 Because of 
11 An analysis of the syntax and style of Smyth’s paratext by Raphael Sackmann and myself is forthcoming in 
Zeitschrift für Celtische Philologie. The question of whether specific stylistic conventions existed for different 
parts of texts needs to be reserved for further comparative research. I can offer here Weiser’s observations on 
the style of John Jewel’s English sermons: he points out that Jewel employed a specific style for the beginning 
of his sermons, which he describes as ‘the rather complex, Ciceronian syntax which in Jewel’s time was thought 
to be the sure sign of a learned man’ (Weiser 1973: 79) – according to Weiser, parallelism has ‘but a small place 
in the Ciceronian style’ (Weiser 1973: 18), it is employed specifically in what Weiser (1973: 121) calls the high 
persuasive style of emotional appeal in Jewel’s sermons.
206
uncertainties of the demarcation of Early Modern ‘sentences’, sentence length is a nu-
merically less reliable ‘arithmetic’ criterion than it appears to be. Other features suscep-
tible to arithmetic analysis, for example the number of constituents in a sentence, will 
require future testing. However, results may potentially hide other significant micro-
stylistic characteristics of individual sentences, and this approach was therefore com-
plemented by a qualitative micro-stylistic analysis. This set out to find features which 
in a second step might then be identified as frequent and pervasive in the text, and thus 
as stylistically significant. At this point, the impressionistic and the arithmetic neces-
sarily and productively overlap. In the specific case under scrutiny here, synonymia, 
parison, and a wide separation of closely related elements by stylistically expanded 
phrases emerged as recurrent devices. Synonymia is a distinctive feature of Smyth’s 
paratext.12 More importantly, synonymia and parison have been described by Adamson 
(2007) and McDonald (2007) as pervasive features of early-modern English writing, 
so by using these figures Robert inscribes himself into a contemporary paradigm of an 
aesthetically valued and prestigious discourse.
Not much fine-grained information is currently available about stylistic features 
of Early Modern Welsh prose works. In order to eventually arrive at a map of rel-
evant features, this paper argues for a text-by-text bottom-up procedure which builds 
on the analysis of individual texts, or text samples, and combines quantitative and 
‘impressionistic’ interpretative perspectives in order to identify notable recurrent mi-
cro-stylistic traits. It reflects on criteria for a stylistic analysis and on difficulties of 
their application, and it highlights a range of options Robert had to structure his text 
within a culturally transmitted set of expectations and norms. More general issues at 
the back of this article concern the applicability of the label ‘Ciceronian’ and the un-
derstanding of dominant modes and models of prose writing in Early Modern Wales. 
Due to the small textual corpus on which it is based, it is very much a methodological 
exercise in empirical historical stylistics, an invitation to apply and develop the cri-
teria suggested here, and to detect further distinctive stylistic traits of Early Modern 
Welsh prose. Historical stylistics is a research area which has much to offer for its 
understanding. Gruffydd Robert, for example, attempted to expand the functional 
and stylistic range of Welsh against a background of Renaissance and humanist ideas 
regarding the advancement of the vernacular; he was aware of foreign-language mod-
els for a refined style of Welsh prose, if, as is likely, we are correct in accepting his 
translation of Cicero as intended as a model, and these larger concerns are arguably 
reflected in the minutiae of his stylistic practice, even when he writes an introduction 
to a catechism.
12 It is, for example, also a feature of Pierre Boaistuau’s Théâtre du Monde, ou il est faict un ample discours des 
miseres humaines and of his Bref discours de l’excellence et dignité de l’homme, of their Welsh translation by 
Roger Smyth as Gorsedd y byd (1615), and of Smyth’s paratexts to the translation.
207
References 
Primary sources
[CLYNNOG, Morys] 1568 Athravaeth gristnogavl, ḷe cair ụedi cynnụys yn grynno’r 
hoḷ brifbynciau syḍ i gyfarụyḍo dyn ar y phorḍ i baradụys. Milan. [digital copy 
available on the ProQuest platform, https://www.proquest.com]
GRUFFYDD ROBERT (1939) Gramadeg Cymraeg gan Gruffydd Robert. Yn ôl yr 
argraffiad y dechreuwyd ei gyhoeddi ym Milan yn 1567. Ed. by G. J. Williams. 
Caerdydd: Gwasg Prifysgol Cymru. [digital copy of the original edition available 
on the ProQuest platform, https://www.proquest.com]
JONES, Howell Ll. Jones/E. I. ROwLANDS (eds) (1975). Gwaith Iorwerth Fynglwyd, 
Caerdydd: Gwasg Prifysgol Cymru. LEWIS, Henry 1948 Hen Gyflwyniadau. 
Caerdydd: Gwasg Prifysgol Cymru.
SMYTH, Rosier (1609) Crynnodeb o aḍysc Cristnogaụl, a Dosparth catholic […], 
Paris. http://hdl.handle.net/10107/5440771
Secondary sources
ADAMSON, Sylvia (2007) “Synonymia: or, in other words.” In: S. Adamson/G. 
Alexander/K. Ettenhuber (eds), 17–35.
ADAMSON, Sylvia/Gavin ALEXANDER/Katrin ETTENHUBER (eds) (2007) Re-
naissance Figures of Speech. Cambridge: Cambridge University Press.
BIBER, Douglas/Susan CONRAD (2019) Register, Genre, and Style. Cambridge: 
Cambridge University Press.
BOWEN, Geraint (1999) Welsh Recusant Writing. Cardiff: University of Wales.
BRYANT-QUINN, Paul (1998) ‘Cymaint Serch i Gymru’: Gruffydd Robert Morys 
Clynnog a’r Athrawaeth Gristnowgawl (1568). Aberystwyth: Canolfan Uwche-
frydiau Cymreig a Celtaidd Prifysgol Cymru.
BRYANT-QUINN, Paul (2000) “To preserve our language: Gruffydd Robert and Mo-
rys Clynnog.” Journal of Welsh Religious History 8, 17–34.
CROLL, Morris W. (1966) Style, Rhetoric, and Rhythm. Edited by J. Max Patrick/
Robert O. Evans with John M. Wallace/R. J. Schoeck. Princeton, NJ: Princeton 
University Press.
DAVIES, Ceri (1995) Welsh Literature and the Classical Tradition. Cardiff: University 
of Wales Press.
GRIFFITH, T. Gwynfor (1953–58) “Italian humanism and Welsh prose.” Yorkshire 
Celtic Studies 6, 1–26.
GRIFFITH, T. Gwynfor (1966) “Italy and Wales.” Transactions of the Honourable 
Society of Cymmrodorion 75/2, 281–298.
GUILLORY, John (2017) “Mercury’s Words: The End of Rhetoric and the Beginning 
of Prose.” Representations 138, 59–86. 
HOFMANN, J.B./Anton SZANTYR (1972) Lateinische Syntax und Stilistik. München: 
C.H. Beck.
208
KÖHLER, Reinhard/Gabriel ALTMANN/Rajmund PIOTROWSKI (eds) (2005) 
Quantitative Linguistik: ein internationales Handbuch – Quantitative linguistics: 
an international handbook. Berlin: de Gruyter.
LORIAN, Alexander (1973) Tendences stylistiques dans la prose narrative française 
au XVIe siècle. Paris: Éditions Klincksieck.
MARSH, David (2013) “Cicero in the Renaissance.” In: Catherine Steele (ed.), The 
Cambridge Companion to Cicero. Cambridge: Cambridge University Press, 
306–317.
McDONALD, Russ (2007) “Compar or parison: or, measure for measure.” In: S. 
Adamson/G. Alexander/K. Ettenhuber (eds), 39–58.
MEHLER, Alexander (2005) “Eigenschaften der textuellen Einheiten und Systeme.” 
In: R. Köhler/G. Altmann/R. Piotrowski (eds), 325–340.
MONFASANI, John (1999) “The Ciceronian controversy.” In: Glyn P. Norton (ed.), 
The Cambridge History of Literary Criticism. Cambridge: Cambridge University 
Press, 393–401.
MUELLER, Janel (2007) “Periodos: squaring the circle.” In: S. Adamson/G. 
Alexander/K. Ettenhuber (eds), 61–77.
OSTROWICZ, Philipp (2003) “Parallelismus.” In: Gert Ueding (ed.), Historisches 
Wörterbuch der Rhetorik. Vol. 6. Darmstadt: Wissenschaftliche Buchgesellschaft, 
546–552.
PALLOTTI, Gabriele (2015) “A simple view of linguistic complexity”. Second Lan-
guage Research 31/1, 117–134.
PARINA, Elena/Erich POPPE (2021) “‘In the Most Common and Familiar Speech 
among the Welsh.’” In: Regine Toepfer/Peter Burschel/Jörg Wesche (eds), Über-
setzen in der Frühen Neuzeit – Konzepte und Methoden/Concepts and Practices of 
Translation in the Early Modern Period. Berlin, Heidelberg: J. B. Metzler, 79–100. 
POPPE, Erich (2019) “Beyond ‘word-for-word’: Gruffudd Bola and Robert Gwyn on 
translating into Welsh.” Studia Celtica Fennica 16, 71–89. 
POPPE, Erich (2022) “How much syntactic complexity could sixteenth-century Welsh 
cope with? The case of Maurice Kyffin’s Deffynniad Ffydd Eglwys Loegr (1595).” 
Zeitschrift für celtische Philologie 69/1, 227–260.
PRICE, Angharad (2019) “Welsh Humanism After 1536.” In: Geraint Evans/Helen 
Fulton (eds), The Cambridge History of Welsh Literature. Cambridge: Cambridge 
University Press, 176–193.
RESCHER, Nicolas (1998) Complexity: A Philosophical Overview. New Brunswick/
London: Transaction Publishers.
RINAS, Karsten (2022) “Vom genormten Satzbau zur genormten Interpunktion. Zur 
Funktion der Zeichensetzung in älterer und neuerer Zeit.” In: Paul Rössler/Peter 
Besl/Anna Saller (eds), Vergleichende Interpunktion – Comparative Punctuation. 
Berlin/Boston: De Gruyter, 109–134.
ROBERT, Jörg (2011) “Die Ciceronianismus-Debatte.” In: Herbert Jaumann (ed.), 
Diskurse der Gelehrtenkultur in der Frühen Neuzeit. Ein Handbuch. Berlin/New 
York: De Gruyter, 1–54.
209
ROBINSON, Ian (1998) The Establishment of Modern English Prose in the Reforma-
tion and the Enlightenment. Cambridge: Cambridge University Press.
SOWINSKI, Bernhard (1999) Stilistik. Stuttgart/Weimar: J.B. Metzler.
STOLT, Birgit (1990) “Redeglieder, Informationseinheiten: Cola und commata in Lu-
thers Syntax.” In: Anne Betten/Claudia M. Riel (eds), Neuere Forschungen zu his-
torischen Syntax des Deutschen. Referate der internationalen Fachkonferenz Eich-
stätt 1989. Tübingen: Max Niemeyer, 379–392. 
THOMAS, Peter Wynn (1996) Gramadeg y Gymraeg. Caerdydd: Gwasg Prifysgol 
Cymru.
TULDAVA, Juhan (2005) “Stylistics, author identification.” In: R. Köhler/G. Altmann/R. 
Piotrowski (eds), 368–387.
WEISER, Daniel K. (1973) The Prose Style of John Jewel. Salzburg: Institut für Eng-
lische Sprache und Literatur, Universität Salzburg.
wESCHE, Jörg (2015) “Stilistik.“ In: Rüdiger Zymner (ed.), Handbuch Literarische 
Rhetorik. Berlin: de Gruyter, 381–398.
wILLIAMS, G. J./Paul BRYANT-QUINN (2019) “ROBERT, GRUFFYDD (c. 1527 
- 1598), priest, grammarian and poet”. Dictionary of Welsh Biography. https://biog-
raphy.wales/article/s12–ROBE-GRU-1522.
Appendix: Corpus ([Clynnog] 1568: [ii]-[vi])
Gruphyḍ fab Rhobert yn annerch yr hyparch brelad, ai ḍibal gynheiliad M. Morys Clynoc: ag yn 
erchi iḍo gan ḍuụ, gynnyḍ, ras a dedụḍụch enaid, a chorph
Gruffydd Robert greeting the venerable prelate and his constant patron, M. Morris Clynnoc, and 
asking for him from God blessing, grace, and felicity of soul and body
(1) [11] VEdi ymy ḍarlain ych ḷyfr, o‘r athraụaeth Gristnogaụl, a chanfod ynḍo megis egin pob 
pụnc hyles i gristion ụrtho, i gadụ‘r enaid, [2] a ḍarfu i ḍuụ i ụneuthur ar i lun, ai ḍelụ: ag a 
rybrynnoḍ Crist ai ụerthfaụr ụaed: [0] e laụenychoḍ fynghalon [12] ụrth ụeled tryssor mor 
ụ[e]rthfaụr yn yr iaith gymraeg: [X] a maint [X11] syḍ o eissie cyfrụiḍid ar phorḍ Grist, yn 
gyphredinol ymysc gụyr yn gụlad: [X] a‘r plant yn crio am fara ([Y] mal y mae‘r prophụyd 
yn ḷefain) [X] heb fod neb, [X12] ai tyrr iḍynt ag ai rhyḍ heb i ụenụyno
After I had read your book on Christian doctrine and found in it as it were in a nutshell 
every point necessary in order for a Christian to preserve the soul, which God has made 
in His image and his likeness and which Christ bought with His precious blood, my heart 
rejoiced to see such a precious treasure in the Welsh language: considering how great is 
the general need for guidance in the way of Christ among the men of our country, and the 
children crying out for bread (as the prophet exclaims) while there is none who breaks it for 
them and gives it without poisoning it.
(2) [0] Am hynny [11] gen ḍarfod i chụi gasclu yn grynno, a dosparth yn drefnus, ag yn eglur 
cymaint o flodeuau, a phynciau iachusaụl, i hyphorḍi vn [21] a chụennychai ụybod sụyḍ, a 
rhann Cristion perpheithgred, [11] i ḍyscu [22] beth a enniḷ nef, beth a dafl ḍyn i vphern, beth 
210
a rhynga boḍ i ḍuụ, a pheth a ụna iḍo sorri: [X] brynti pechod, odidoụgrụyḍ rhinụeḍ: [0] 
ni eḷais ar fynghalon [12] na pharụn i brintio: [23] fal y gaḷo eraiḷ [31] syḍ ag eissie y cyfryụ 
ymborth ysprydol arnynt, [23] fod yn gyfrannol o‘r ụleḍ [32] a ḍarfu ichụi i harlụy. 
Therefore, because you have assembled compactly and arranged orderly and clearly so 
many flowers and wholesome articles in order to instruct the one who would wish to know 
the duty and the share of a Christian of perfect religion, to teach what will win heaven, what 
will cast man to hell, what will please God, and what will make him displeased: pollution 
of sin, excellence of virtue: I did not have the heart not to bring about its printing so that 
others who lack such spiritual sustainance could partake in the feast you have prepared.
(3) [0] Gobeithio [2] pan ḍelo i ḍụylaụ y crefyḍgar gymru, [1] y gụna laụer o les iḍynt, trụy i 
hụylio i baradụys, ai troi o phorḍ uphernaụl.
Let’s hope that when it comes into the hands of the pious Welsh, it will do them much good, 
by directing them to paradise and turning them from the infernal road.
(4) [0] E fyḍ tostur fynghalon [1] pan feḍyliụyf faint [2] syḍ o blant trụy dir cymru, odidaụg i 
athrylith, a darpar gụyr arḍerchaụg, yn methu ag yn cymryd ḷụybr annụyiaụl eisiau cael oi 
mebyd i hypho[r]ḍi meụn dysc, ai meithrin meụn moessaụl gampau. 
It wounds my heart when I think how many children throughout Wales, magnificently talen-
ted and potentially splendid people, fail and take the road of ungodliness lacking to get from 
their youth guidance in doctrine and education in moral development.
(5) [01] Yr achos fụyaf o hynn yụ diphig ḷyfrau [1] a draethant o‘r cyphelib ystyr (.) [02] ond 
yroụron e ḍarfu i chụi meụn ychydig o ḍolennau roḍi cymorth, a help iḍynt rhag yr eissiau 
hynn 
The foremost reason for this is a want of books which set out such contents, but now you 
have given them in a few pages succour and help against this deficiency.
(6) [0] Canys yn ych ḷyfr chụi yma nhụy a gaant oi ḍysgu yn haụd, meụn ychydig o amser, a 
thrụy ychydig help, a ḷai o gost, y pethau [1] syḍ angenrheidiol iụ gụybod, i hen ag ifanc.
Since in this book of yours they will find, to teach them easily, in little time and with little 
help and less cost, the things that are necessary to know for old and young.
(7) [0] Canys pụy yụ hụnnụ [1] a eiḷ ḍoedyd [21] i fod yn gristion, [22] oni ụyr [3] pa foḍ y mae 
credu yngrhist, beth syḍ oi obeithio gentho, a pheth a orchmynoḍ ef i gadụ; beth a ụaharḍoḍ 
i ụneuthur, beth a ennil obrụy, a pheth a hauda gosp?
Since who is the one who can say that he is a Christian if he does not know how one believes 
in Christ, what is expected by him, and what he commanded to keep, what he forbade to do, 
what wins reward and what incurs punishment?
(8) [0] fely [11] pan ystyrio‘r cymru [22] syḍ yn caru i heneidiau, [23] mor anhepcor ydyụ‘r 
rhain, ag mor hauḍ i dyscu ụrth ḍarlain y traethiad yma: [0] nhụy a ‘madaụant ai diogsụrth 
211
eisteḍach, ai bustlaụl sertheḍ, ag ai smala gyfeḍach ([12] onid ydynt ụedi boḍi meụn brynti 
pechod) [0] ag a ‘mroḍant i ḍyscu pethau sprydol, buḍfaur i‘r enaid.
Thus, when the Welsh who love their souls, contemplate how necessary these are and how 
easy to learn by reading this treatise, they will renounce their tardy lolling about and their 
foul obscenity and their vain merriment (unless they have been submerged in pollution of 
sin) and apply themselves to learn spiritual things, beneficial for the soul.
(9) [0] A hynn nis caant meụn mann araḷ yn y byd mor fyrr, mor drefnus, mor eglur oi deaḷt ag 
yn y ḷyfr yma i chụi. 
And these [i.e., the spiritual things], they will not find them in any other place at all as brief-
ly, as orderly, as clearly to be understood as in this book of yours.
(10) [0] canys amhossibl oeḍ gynnụys meụn ḷai o erriau, a dosparth yn oleuach, a chyfleu yn 
ụeḍeiḍiach gynnifer bynciau, a chyn ḍyfned i ‘styriaeth; [1] fal y gaḷo y plant a‘r gụrageḍ i 
deaḷt, 
For it would be impossible to contain in fewer words, and to structure more clearly, and to 
arrange more suitably, so many subjects – and so deep their meaning – so that the children 
and the women may understand them.
(11) [0] gụyn i byd trụy gymru, [1] pe parent ymhob eglụys ụrth aros y gụasanaeth, ne ar osteg 
ypheren, gartref ymysc tylụyth y ty i difyrru‘r amser ag ymhob cyniḷeidfa i ḍiḍanu‘r bobl, 
ḍarlain hụnn ne‘r cyfryụ ymadrodion a gadel i phorḍ henchụedlau coegion, a choụydau 
gụenheuthus, celụḍog.
It would be a great blessing throughout all Wales if they [the Welsh people] made a habit, 
in every church while waiting for the service to begin, or during low Mass, at home among 
the household to shorten the time, and in every assembly to comfort the people, of reading 
this book or similar material and have done with old, false legends and flattering, lying 
cywyddau.
(12) [0] Ond [11] ar hydr y rhyḍ yr yspryd glan ras iḍynt hụy i gymryd dysc, [21] megis y roes i 
chụi oi scrifennu attyn; [0] mi danfonaf yrhain yn i mysc, [12] dan erfyn ar ḍuụ ymhob gụeḍi 
[22] a ụnelụyf [23] ar ụrtheithio hono i calonnau nhụy i ḍerbyn aḍysc, a rhoi nerth i chụithau 
i scrifennu chụaneg er ḷes i‘r Gristnogion, a gogoniant i duụ.
But in the hope that the Holy Ghost will give them grace to accept teaching, as it gave it to 
you to write to them, I will send these to them, entreating God in every prayer I pray that he 
equips their hearts to receive instruction and that he gives strength to you to write more for 
the benefit of Christians and God’s glory.
212
Abstract
BETWEEN THE IMPRESSIONISTIC AND THE ARITHMETIC:
THINKING ABOUT CRITERIA FOR THE STYLISTIC ANALYSIS OF EARLY 
MODERN WELSH PROSE
Empirical historical stylistics is methodologically a difficult field since it ‘perhaps 
inevitably combines the impressionistic and the arithmetic’ (Guillory 2017: 63). For 
lesser researched languages or periods, the problems are aggravated because even im-
pressionistic assessments on which further hypotheses and comparative work could 
be built, are rare. Early Modern Welsh (c. 1500 – c. 1700) is a period to which this 
qualification applies. This article will discuss some methodological issues and param-
eters for a micro-stylistic analysis of Early Modern Welsh prose, i.e., on the level of 
individual sentences. Its approach is bottom-up, taking as its point of departure the in-
troductory paratext to Morys Clynnog’s catechism Athravaeth Gristnogawl (‘Christian 
Doctrine’, 1568) by its editor Gruffydd Robert. It argues that in the case of lesser re-
search languages, empirical historical stylistics will need to proceed from the analyses 
of individual texts or text samples which combine quantitative and ‘impressionistic’ 
interpretative perspectives in order to identify notable recurrent micro-stylistic traits.
Keywords: empirical historical stylistics, Early Modern Welsh, Gruffydd Robert, 
Ciceronianism
Povzetek
MED IMPRESIONISTIČNIM IN ARITMETIČNIM: KAKŠNI NAJ BI BILI KRI-
TERIJI ZA STILISTIČNO ANALIZO ZGODNJE MODERNE VALIŽANSKE 
PROZE
Empirična historična stilistika je v metodološkem pogledu težavno področje, glede na 
to, da se tu “morda neizogibno srečujeta impresionistično in aritmetično” (Guillory 
2017: 63). Pri manj raziskanih jezikih ali obdobjih je težava še večja, saj so v zvezi 
z njimi redke celo impresionistične ocene, na katerih bi lahko temeljile nadaljnje hi-
poteze in primerjalne študije. Takšen primer je zgodnja moderna valižanščina (pribl. 
1500 – pribl. 1700). Pričujoči članek obravnava nekaj metodoloških vprašanj in para-
metrov za mikrostilistično analizo zgodnje moderne valižanske proze, in sicer na ravni 
posameznih povedi. Uporabljen je pristop “od spodaj navzgor”, kot izhodišče pa služi 
uvodni paratekst h katekizmu Athravaeth Gristnogawl (‘Krščanska doktrina’, 1568) 
Morysa Clynnoga, ki ga je napisal urednik Gruffydd Robert. Članek skuša pokazati, da 
bo pri manj raziskanih jezikih empirična historična stilistika morala najprej analizirati 
posamezna besedila ali odlomke besedil in opazovane značilnosti razložiti ob upošte-
vanju tako kvantitativne kot “impresionistične” perspektive, s ciljem prepoznavanja 
pomembnih ponavljajočih se mikrostilističnih značilnosti.
Ključne besede: empirična historična stilistika, zgodnja moderna valižanščina, Gru-
ffydd Robert, ciceronizem