5 4 Jana Le ická Usage and Empi ical P oduc i i y o In e na ional
Adjec i al Su ixes in Slo ak Re isi ed
doi.o g/10.35321/ e m28-03
Usage and Empi ical P oduc i i y
o In e na ional Adjec i al Su ixes
in Slo ak Re isi ed
Jana Le ická
Ľ. Š ú Ins i u e o Linguis ics, Slo ak Academy o Sciences
ORCID id: h ps://o cid.o g/0000-0001-6027-604X
ABSTRACT
This pape ep esen s a sequel o ea lie esea ch ocusing on he usage and
p oduc i i y o i e La ina e su ixes in Slo ak. The analysis ocuses on eal and
po en ial p oduc i i y in a wo-s age compa ison: 1) okens and lemmas occu -
ing in a gene al balanced co pus s gene al co pus o specialised and academic
ex s, 2) gene al co pus o specialised and academic ex s s specialised (sub)-
co po a o medical, legal, economic and eligious ex s. The i s aim o he se-
quel was o es ablish i he low- equency lemmas include new coinages ha
could con ibu e o he p oduc i i y a ia ion ac oss domains. The second aim
was o iden i y and implemen an app op ia e s a is ical measu e ha would en-
able he compa ison o p oduc i i y co pus da a ac oss di e en ly sized co po a.
KEYWORDS: p oduc i i y, adjec i e, su ix, gene al co pus, specialised co pus, hapax lego-
mena, low- equency lemma, la ge numbe o a e e en s dis ibu ions.
ANOTACIJA
Šiame s aipsnyje ap ašomas y imas y a anks esnio y imo, nag inėjančio penkių
lo ynų kalbos p iesagų a ojimą i p oduk y umą slo akų kalboje, ęsinys. Ty ime
siekiama nus a y i ealų i po encialų p oduk y umą a liekan palyginimą d iem
e apais: 1) žodžių o mos (angl. okens) i an aš inės žodžių o mos, a ba lemos
(angl. lemmas), esančios bend ajame subalansuo ame eks yne, lyginamos su
esančiomis bend ajame specialiųjų i akademinių eks ų eks yne; 2) lyginamos
bend ajame specialiųjų i akademinių eks ų eks yne i specialiajame medicininių,
eisinių, ekonominių i eliginių eks ų pa eks ynyje esančios žodžių o mos i
lemos. Pag indinis ęs inio y imo ikslas – nus a y i, a a p e ai a ojamų lemų
y a naujažodžių, ku ie galė ų p isidė i p ie p oduk y umo ski umų ski ingose s i-
yse. Ty imu aip pa siekiama nus a y i i p i aiky i s a is inę p iemonę, ku i leis ų
palygin i eks yno medžiagos p oduk y umą ski ingos apim ies eks ynuose.
ESMINIAI ŽODŽIAI: p oduk y umas, būd a dis, p iesaga, bend asis eks ynas, specialusis ek-
s ynas, hapaksas, e ai a ojama lema, didelis e ų į ykių pasiski s ymų skaičius
5 5Te minologija | 2021 | 28
1. INTRODUCTION
The mo i a ion o ou esea ch a ose om an in e es in e minologi-
cal synonymy and, wi hin his b oade opic, in synonymous denominal
adjec i es as cons i uen s o mul i-wo d e ms.
The ocus on adjec i es is no a no el y in e minology esea ch. De-
ailed analyses aimed a adjec i es can be ound in he wo k o a numbe
o e minologis s, especially om he 1990s and 2000s1. In he his o y o
Slo ak e minology, adjec i e-cen ed s udies, igge ed by he need o
uni y and coin s anda dised Slo ak e minologies, can be aced o he
1950s (see e.g. Ján Ho ecký 1956).
The impo ance o adjec i es in e minology and languages o specialized
pu poses can also be illus a ed by s a is ical da a p o ided by co po a. The
majo i y o specialised co po a ha eme ged om he Slo ak Na ional Co -
pus p ojec (he eina e only SNC p ojec ) ea u es a highe a io o adjec-
i es compa ed o he e e ence co pus (p im-7.0- k), i.e., he co pus com-
piled o an e en sha e o ic ional, specialised and jou nalis ic ex s2. As can
be seen in Table 1, his di e ence amoun s o almos 2%, excep o he
eligious co pus (bl -2.0) in which he a io o adjec i es is oughly he same
as in he e e ence co pus (7.46%). The second ow o Table 1 p esen s he
a io o ge unds ha a e also used wi h di e en ia ing and classi ying unc-
ions in Slo ak mul i-wo d e ms. In he amewo k o he mo phological
anno a ion o he SNC p ojec , hese a e classi ied as a speci ic g oup, which
is why hey a e agged sepa a ely om e bs (Ga abík e al. 2004).
The coexis ence and compe i ion o adjec i es has been iden i ied and
analysed in he con ex o language in gene al (e.g. Nábělko á 1996).
Howe e , hei en e ing in o mul i-wo d e ms ep esen s a a he unde -
esea ched opic. A common phenomenon esul ing in synonymy and
compe i ion o adjec i es is he coexis ence o La ina e adjec i es and hei
(Slo ak) coun e pa s ( nú ožilo ý – in a enózny “in a enous”). The
la e g oup comp ises also a subg oup ha ea u es bo h Slo ak and, om
he e ymological poin o iew, La ina e o in e na ional su ix combined
wi h he same in e na ional oo (e.g. bak e iálny – bak é io ý “bac e ial”).
Occasionally, a a ia ion o wo in e na ional su ixes wi h he same in-
1 See, o example, H. Assadi and D. Bou igaul 1995; S. No mand and D. Bou igaul 2001; B. Daille
2001; F. Maniez 2001; M.-C. L’Homme 2002, 2003; o I. Ca iè e 2008.
2 Mo e de ails conce ning he co po a used o p esen ed analysis can be ound in Pa 3.2.
5 6 Jana Le ická Usage and Empi ical P oduc i i y o In e na ional
Adjec i al Su ixes in Slo ak Re isi ed
e na ional oo can a ise (e.g. hype sonó ny – hype sonický, kon adik-
o ický – kon adik ó ny).
This s udy, ocusing on he usage o speci ic su ixes in he Slo ak
language, seeks o iden i y whe he some o hem a e mo e likely o a-
cili a e such a coexis ence o o occu in e minological adjec i es mo e
o en, which is why he concep o mo phological p oduc i i y becomes
ele an in his case.
The su ixes selec ed o he analysis ep esen a mino g oup used o
coining denominal adjec i es wi hin a wide ange o Slo ak adjec i al
su ixes3. As o hei composi ion, he i e su ixes consis o an (adap -
ed) in e na ional adjec i al su ix combined wi h seman ically equi alen
Slo ak -ný/ny: -álny, -á ny, -i ný, -í ny, -ózny, hus, hei adjec i al
s a us is somewha mul iplied. I appea s ob ious ha all i e su ixes o
La in o igin we e used o coin adjec i es om La in nouns (-alis wi h
he a ian -a is, -ī us, -ōsus) o e bs (-ō ius). I is wo h poin ing ou ,
howe e , ha many Slo ak adjec i es wi h he analysed su ixes en e ed
3 Ma in Ološ iak and Lucia Ološ iako á (2015: 230) men ion as many as 38 su ixes, hough 25 su ixes
in hei sample – de i ed om he Slo ník ko eňo ých mo ém slo enčiny comp ising 66 500 lexical
uni s – ep esen only 1% o adjec i es.
Table 1. Numbe o okens and a ios o adjec i es and ge unds in he e e ence
and specialised co po a o he SNC p ojec
COR-
PUS
eFe -
ENCE
CORPUS
SCIeNTIF-
IC SUB-
CORPUS
MeDICAL
SUB-
CORPUS
LEGAL
SUB-
CORPUS
RELI-
GIOUS
CORPUS
ECO-
NOMIC
CORPUS
Numbe
o okens
in he
co pus
253,137,609 149,581,785 7,099,555 33,600,183 65,920,357 164,987,015
Numbe
o adjec-
i es in
okens
19,090,396
7.54%
13,415,554
8.97%
660,025
9.3%
4,841,400
9.88%
4,914,860
7.46%
15,540,381
9.42%
Numbe
o ge -
unds in
okens
3,174,567
1.25%
2,394,914
1.6%
112,164
1.58%
1,267,598
2.59%
761,719
1.16%
2,154,124
1.31%
5 7Te minologija | 2021 | 28
he Slo ak lexicon no only di ec ly om La in, bu mo e o en ia F ench
o English in he las decades (Ho ecký 1999) and unde wen he p ocess
o adap a ion by means o he a o emen ioned Slo ak adjec i al su ix. I
is signi ican ha his bo owing and subsequen adap a ion was enabled
and enhanced by ex a-linguis ic ac o s. Due o p edominan ly his o ical
conside a ions, Slo ak has been e y open o new wo ds o igina ing in
La in and hus he sha e o La ina e wo ds is high, e en in he gene al
lexicon o Slo ak (Ho ecký 1999: 81). Nowadays, hanks o he impac o
mass media, in e ne , and social media, his combined wo d- o ming
p ocess o bo h bo owing and de i a ion is exceedingly p o i able and
hus highly pe inen o a linguis ic and e minological s udy.
Ou pape ep esen s a sequel o ou p e ious esea ch on he same
esea ch opic, which is cu en ly in p in (Le ická 2021). Fo he sake
o compa ison and cla i y o he eade , we conside ed i necessa y o
include a summa y o he p e ious easoning and indings, supplemen ed
wi h new insigh s and heo e ical backg ound. Thus, some o he ables
(1–4), hough hey appea ed in ou i s pape , a e epea ed he e.
2. THEORETICAL UNDERPINNING
The heo e ical concep o mo phological p oduc i i y has been de eloped
by qui e a numbe o esea che s and schola s. An o e iew o p e ious
heo e ical hinking and s udies in he 20 h cen u y has been summa ized
and commen ed on in a numbe o s udies and a icles, e.g. by Jesús Fe nán-
dez-Dominguez (2013) o Vic o ia Hulse (2011). He e, we will ocus only
on a numbe o heo e ical conside a ions, which bo h p o ided us wi h
inspi a ion and in luenced ou choice o he app oach and me hodology.
The oldes de ini ion quo ed he e was w i en in 1948 by Ame ican
linguis Dwigh L. Bolinge (quo ed by Säily 2018: 198), who concluded
ha p oduc i i y was he “s a is ically de e minable eadiness4 wi h
which an elemen en e s in o new combina ions”.
In 1961, Du ch linguis Henk Schul ing (quo ed by S e an E e and
Anke Lüdeling 2001: 165) de ined p oduc i i y as “ he possibili y o
language use s o coin unin en ionally and, in p inciple, unlimi ed
numbe s o new o ma ions, by using he mo phological p ocedu e ha
lies behind he o m-meaning co espondence o some known wo ds”.
4 The ph ases o wo ds ma ked in bold le e s ep esen he key ea u es o ou esea ch.
5 8 Jana Le ická Usage and Empi ical P oduc i i y o In e na ional
Adjec i al Su ixes in Slo ak Re isi ed
Las , bu no leas , we would like o e e o he wo d- o ma ion esea ch
o Czech linguis Miloš Dokulil, who p esen ed a complex heo y o wo d-
o ma ion p oduc i i y in his wo k Teo ie oření slo in 1962. Fi s o all,
Dokulil de ined he p oduc i i y o a language de ice in gene al as an “abil-
i y o his language de ice (be i a base, a ix o wo d- o ma ion pa e n) o
ac i ely pa icipa e in coining new wo ds” (1962: 78). He speci ied ha his
abili y can be absolu e, i.e. i s usabili y in coining new wo ds in any ime and
wi hin any wo d- o ma ion pa e n, o ela i e which is seman ically, o
a he s yle dependen as well as wo d- o ma ion pa e n dependen . In
Dokulil’s heo y, he p oduc i i y is a “concep o synch onic na u e, while
he implemen a ion o his abili y – coining new wo ds acco ding o spe-
ci ic ules – is, o he con a y, a concep o diach onic na u e” (1962: 80).
Dokulil also di e en ia ed be ween sys emic and empi ical p oduc i i y (also
e med “pa ole” o “ eal p oduc i i y” by F amn išek Š ícha 2012)), which
gi es a “gene al idea o he o e all exploi a ion o a speci ic wo d- o ma-
ion pa e n o language de ice in he sys em o a language in a gi en ime”
(Dokulil 1962: 80). Dokulil belie ed ha e en “app oxima i e da a con-
ce ning he quan i a i e use o a gi en wo d- o ma ion p ocess o elemen
a e o pa amoun impo ance o he o e all pic u e o a gi en language in
gene al and o he cha ac e is ics o i s lexicon in pa icula ” (I em: 77).
F om among con empo a y Czech linguis s building on Dokulil’s he-
o y o p oduc i i y and e i ying his assump ions on co pus da a (see e.g.
Š ícha 2002, 2007, 2009, 2012 o Magda Še číko á 2014), Š ícha sugges s
he analysis o empi ical/pa ole p oduc i i y, no only in big co po a, bu
also by means o a se ies o co po a o di e en sizes and composi ion
(2012: 104).
I could be claimed ha Dokulil’s assump ions conce ning he ela i e
p oduc i i y we e echoed by Ha ald Baayen and o he s whose me hods
and app oaches we will e e o la e in his analysis. In 1999, Baayen and
his colleagues Ingo Plag and Ch is iane Dal on-Pu e w o e ha “claims
abou he p oduc i i y o a gi en a ix a e gene ally made wi hou di -
e en ia ing p oduc i i y acco ding o he ype o discou se, al hough i
is commonly assumed ha ce ain kinds o de i a ional su ixes a e mo e
pe inen in ce ain kinds o ex s han in o he s” (1999: 209).
Wi h la ge co po a a ailable, i is possible o measu e, iden i y and
analyse he p oduc i i y o a wo d- o ma ion pa e n o elemen no on-
ly in gene al language a a gi en ime, bu also in specialised domains
and egis e s.

5 9Te minologija | 2021 | 28
3. PREVIOUS RESEARCH
In line wi h he easoning and assump ions o linguis s quo ed in he
p e ious pa , we o mula ed h ee key esea ch ques ions a he onse o
ou esea ch aimed a iden i ying he mo phological p oduc i i y o ad-
jec i al su ixes:
1) Is he mo phological p oduc i i y o indi idual adjec i al su ixes
om he selec ed g oup di e en in gene al and specialised ex s?
And a e hese di e ences s a is ically signi ican ?
2) Can he mo phological p oduc i i y o indi idual adjec i al su ix-
es om he selec ed g oup a y depending on he speci ic do-
main? A e some o hose i e su ixes mo e “ap ” o con ibu e o
he encoding o ield o domain speci ic concep s?
3) Wha me hods a e ele an o his speci ic analysis o co pus da a?
Fi s o all, due o he speci ici y o his kind o wo d- o ma ion pa e n
La in base + La in su ix + e ymologically na i e su ix, we decided o se
aside he quali a i e analysis in ol ing he limi s and ules go e ning he
combina ion o indi idual La in bases wi h one o he selec ed adjec i al
su ixes in he con ex o he Slo ak language.
Al hough he quan i a i e app oach canno p o ide he whole pic u e
o he p oduc i i y s a us o a su ix o any o he mo phological o m,
i s bene i o his ype o esea ch lies in p esen ing he pe spec i e,
he con ex o usage o wo ds coined acco ding o a speci ic pa e n,
along wi h speci ic compa isons wi h seman ically and unc ionally equi -
alen o ms. I is a gene ally accep ed iew ha a quan i a i e app oach
in he analyses o mo phological p oduc i i y ep esen s a signi ican ,
complemen a y me hod o quali a i e esea ch (see e.g. E e and Lüdel-
ing 2001). As Plag s a es: “quan i a i e and quali a i e no ions o p o-
duc i i y […] a e closely ela ed. Thus, he idea o po en iali y, which
is cen al o quali a i e de ini ions o p oduc i i y, can be exp essed in
he s a is ical e ms o p obabili y” (quo ed by Nacca a o 2016: 135).
Simila ly, Baayen and his co-au ho s ad oca e ha p oduc i i y “seems
o be a scala concep […] wi h some a ixes one is mo e likely o en-
coun e newly o med wo ds han wi h o he , a ac ha makes p oduc-
i i y a p obabilis ic no ion which is suscep ible o s a is ical analysis”
(1999: 10).
6 0 Jana Le ická Usage and Empi ical P oduc i i y o In e na ional
Adjec i al Su ixes in Slo ak Re isi ed
3.1. S a is ical me hods
In s age 1 o ou analysis, we adop ed he well- odden me hodological
pa h o Ha ald Baayen and his colleagues and ollowe s (1991, 1992, 1993,
1994, 1996, 1997). We iden i ied he so-called ealized p oduc i i y, i.e.
he equency o usage o adjec i es coined wi h selec ed su ixes as well
as he numbe o di e en adjec i es coined wi h hese su ixes in di e -
en co po a. These co pus s a is ics ou line he p oduc i i y o su ixes
wi h espec o he pas and p esen linguis ic si ua ion.
The nex s ep was o de e mine he po en ial p oduc i i y o he analysed
su ixes, i.e. an es ima e o he a e a which new ypes5 a e expec ed o
appea . Baayen (2009: 7) sugges s calcula ing i as he a io o hapax lego-
mena6 wi h a ix X and all okens wi h a ix X in a co pus. The usage o
hapax legomena o he calcula ion o p oduc i i y has been widely c i i-
cised, as his g oup does no di ec ly ep esen new coinages ha a e sup-
posed o e lec he p oduc i i y o an elemen o pa e n. Plag, Dal on-
Pu e and Baayen a gue ha he po en ial p oduc i i y is a p obabilis ic
measu e, and ha wi h he inc ease in size o a co pus, “ he p opo ion o
neologisms among he hapax legomena inc eases and i has been shown
ha i is p ecisely among he hapax legomena ha he g ea es numbe o
neologisms appea ” (1999: 12). In he p e ious s age o ou esea ch, we
decided agains using he hapax/ oken me hod, bu adop ed he hapax/
ype me hod in line wi h an Ma le’s easoning (1992) ha oken equen-
cy is no as ele an a a iable in a measu e o p oduc i i y as he numbe
o lemmas. Howe e , he mos se ious d awback imposed by his measu e
is he impossibili y o compa ing he s a is ical da a o co po a o di e en
sizes (Nacca a o 2016: 133). We will e u n o his conside a ion in Pa 5.
3.2. Co po a used in he analysis7
All h ee co po a and h ee subco po a used in his analysis we e e-
leased by he Depa men o he SNC in 2013–2020 and a e accessible
o all egis e ed use s.
5 Types in co pus linguis ics ep esen unique wo ds, hey a e synonymous o lemmas (Bake e al. 2006).
6 Hapax legomena (o ig. G eek ph ase meaning “once said”), also abb e ia ed o hapax, is a wo d ha
occu s only once in a pa icula co pus (Bake e al. 2006).
7 Mos o his sec ion comes om he pape desc ibing he p e ious s age o ou esea ch (Le ická 2021),
howe e , his epe i ion is necessa y in o de o explain he na u e o he co pus da a.
6 1Te minologija | 2021 | 28
The i s , e e ence co pus (p im-7.0- k), amoun ing o mo e han 253
million okens, is composed o an e en sha e o jou nalis ic, specialised
and ic ional ex s (64.12% o hem a e o iginally w i en in Slo ak, while
29.51% ep esen ansla ions), w i en in 1991–2015. The co pus was
used in he compila ion o wo equency dic iona ies o Slo ak (2017,
2018) as well as he e e se dic iona y (2018).
The second co pus, p im-9.0-public-p , is a publicly a ailable subco -
pus o he p ima y co pus o he SNC p ojec (he eina e e e ed o as
he scien i ic subco pus). Compiled om specialised, academic and non-
ic ion ex s, his subco pus ea u es mo e han 149 million okens and
documen s, as well as he gene al discou se o science and esea ch, in-
cluding specialised jou nalism. I s ex s we e w i en be ween 1955–2019.
The pe cen age o ex s w i en in 1955–1991 amoun s o 8.7% (mo e
han 13 million o okens).
The smalles (sub) co pus o all he sea ched co po a is he esul o
il e ing he p ima y co pus o he SNC p ojec [9], e sion 9.0. I consis s
o ex s ha belong o he ield o medicine w i en in 1976–2019 and
comp ising sligh ly mo e han 7 million okens. The pe cen age o ex s
w i en in 1955–1991 amoun s o 1.09%. We will e e o i h oughou
he ex as he medical subco pus.
In o de o ha e a compa able sou ce o igina ing om o he specialised
co po a and he e e ence co pus, he specialised co pus legal-1.1 (he e-
ina e e e ed o as he legal subco pus), buil in coope a ion wi h he
Slo ak Minis y o Jus ice, was na owed down o legisla i e ex s c ea ed
in he pe iod 1991–2011. I s almos 49 million okens we e hus educed
o 33.5 million okens.
The specialised co pus – bl -2.0 – ocusing on he eligious domain,
was eleased in 2014. I s ex s consis o almos 66 million okens w i en
be ween 1989–2014 (he eina e e e ed o as he eligious co pus). I
comp ises mo e han 80% o hema ic jou nals and newspape s. Simi-
la ly, specialised co pus ecn-2.0-public (he eina e e e ed o as he
economic co pus), de o ed o he domain o economics, includes as much
as 96.24% o specialised ex s published in hema ic jou nals and news-
pape s. Tex s o his co pus come om 1992–2014 and comp ise almos
165 million okens.
I mus be no ed, howe e , ha hese co po a a e he e ogenous and
co e a wide a ie y o ex s.
6 2 Jana Le ická Usage and Empi ical P oduc i i y o In e na ional
Adjec i al Su ixes in Slo ak Re isi ed
As a as he co pus sea ch is conce ned, we did no base he que ies
on mo phological agging p o ided o each and e e y (sub) co po a in-
cluded in he analysis, on he basis ha , wi h La ina e wo ds, he agging
p o ed o be inadequa e and e oneous. The e o e, we op ed o a simple
sea ch o he speci ic ending o a oken, e.g. [lemma=“.*álny“], combined
wi h he au oma ic il e ing o wo ds wi h inciden ally he same s ing o
cha ac e s (see Pa 4.1).
3.3. Resul s o he p e ious s age
In he p e ious s age, s a is ical da a om co po a enabled us o c ea e
a anking o he ealized p oduc i i y o he analysed su ixes. The su ix
-álny is de ini i ely he mos widely used in all (sub) co po a, while -ó ny
is a he opposi e pole o he equency axis in 5 (sub) co po a. As can be
seen om he no malised equencies o usage o ou ou o i e su -
ixes, hey appea mo e equen ly in specialised and academic ex s (sci-
en i ic subco pus) compa ed o he e e ence co pus. I is also no ewo hy
ha he no malised equencies o su ixes in he medical subco pus ei he
equal o conside ably exceed he ipm in he scien i ic subco pus, while
he ipm o su ixes (ins ances pe million) in he eligious co pus is man-
i es ly lowe han in he e e ence co pus. This can be pa ly explained
by he ype o ex s included in hose (sub) co po a.
Table 2. F equency and no malised equency (IPM, ins ances pe million) o he analysed su ixes
in espec i e (sub) co po a. The co po a a e o de ed om he smalles o he la ges one
SuF-
FIx
MeDICAL
SUB-
CORPUS
LEGAL
SUB-
CORPUS
RELI-
GIOUS
CORPUS
SCIeNTIF-
IC SUB-
CORPUS
ECO-
NOMIC
CORPUS
eFe -
ENCE
CORPUS
Tokens Tokens Tokens Tokens Tokens Tokens
IPM IPM IPM IPM IPM IPM
-álny 31,408 102,863 128,472 537,682 597 005 547,742
4423.94 3061.38 1948.9 3594.57 3618.5 2163.81
-á ny 4,529 25,980 14,253 94,586 38,178 84,780
637.93 773.21 216.22 634.14 231.40 334.92
-i ný 2,683 2,256 9,176 30,402 52,574 44,655
377.91 67.14 139.20 203.25 318.66 176.41
6 9Te minologija | 2021 | 28
a gi en su ix. Compa ed o Table 4, he anking changed in all (sub)-
co po a excep o he e e ence co pus. Thus, i seems ha he inclusion
o low- equency lemmas p o ed o be wo hwhile.
The su ix -ó ny is mos p oduc i e in wo co po a (compa ed o h ee in
Table 4), while -ózny emained a he op o wo ankings and -i ný a he
op o he anking in he economic co pus. In he legal co pus, he mos
p oduc i e su ix seems o be -á ny while his e y su ix is simul aneously
leas p oduc i e in medical ex s and economic ex s. The las place in he
p oduc i i y ankings is occupied by he su ix -ó ny bo h in he e e ence
co pus and medical ex s, by he su ix -ózny in gene al scien i ic ex s and
he su ix -i ný emained leas p oduc i e in legal and eligious co po a.
5. COMPA INg P ODuCTI ITy AC OSS (SuB) CO PO A
As we poin ed ou in Pa 3.1, s a is ical measu es using aw equencies
and coun s o ypes p e en ed us om compa ing p e ious esul s o igi-
na ing om di e en (sub) co po a. I we ag ee wi h he claim o a he
he hypo hesis ha he p oduc i i y o any wo d- o ma ion elemen o
Table 9. Di e ence in a io o po en ial neologisms be ween he g oup o low- equency lemmas
(4–1 occu ences) and he numbe o ypes. The anking o su ixes ha a e highligh ed changed
compa ed o Table 4
LOW-F equeNCy LeMMAS/TyPe ATIO
e e ence co pus scien i ic subco pus medical subco pus
-ózny 0.2689075 -ó ny 0.2727272 -ózny 0.2638888
-á ny 0.2466666 -álny 0.2716468 -álny 0.1682785
-álny 0.2053838 -á ny 0.2712328 -ó ny 0.1538461
-i ný 0.1333333 -i ný 0.2653061 -i ný 0.1388888
-ó ny 0.12 -ózny 0.2436974 -á ny 0.1284916
legal subco pus eligious co pus economic co pus
-á ny 0.1212121 -ó ny 0.3 -i ný 0.3504273
-ó ny 0.1 -álny 0.2651391 -álny 0.2647849
-ózny 0.1 -ózny 0.2115384 -ózny 0.2560975
-álny 0.0776397 -á ny 0.1985815 -ó ny 0.2307692
-i ný 0-i ný 0.1818181 -á ny 0.225

7 0 Jana Le ická Usage and Empi ical P oduc i i y o In e na ional
Adjec i al Su ixes in Slo ak Re isi ed
pa e n can be concluded om new coinages (and, a he same ime, we
p o ed in Pa 4.2 ha hese new coinages could occu no only wi hin
hapax g oups, bu also among low- equency ypes), i is possible o base
u he analysis on wo d- equency dis ibu ions and equency spec a.
By “ equency spec um” we mean a lis epo ing how many ypes in a
equency lis can be obse ed o occu (Ba oni 2009: 807).
F om he pe spec i e o wo d- equency dis ibu ion, e e and Lüdeling
(2001: 166), in line wi h Schul ink’s de ini ion, claim ha “a p oduc i e
pa e n is, in heo y, cha ac e ised by an in ini e ocabula y (c . he no ion
o un-limi edness in Schul ink’s de ini ion), whe eas a o ally unp oduc i e
pa e n is expec ed o ha e a ini e, and o en qui e small, ocabula y”.
They poin ou ha “low equency ypes (including hapax legomena, bu
also ypes occu ing wo, h ee, e c. imes) accoun o a majo pa o he
ocabula y” o p oduc i e pa e ns and conclude ha his kind o com-
pa ison canno be based on s anda d s a is ical models (2001: 167).
Fo his pu pose o s a is ical modelling, Baayen in oduced in 2001 he
so-called LNRE10 dis ibu ions. The bene i o hese specialised models
lies in he possibili y o ex apola ing he ype- oken s a is ics om a
speci ic sample o la ge alues o okens and hus es ima e po en ial
neologisms ou side his sample. They a e based on he p edic ed ocabu-
la y sizes ob ained om a coun o low- equency ypes in he co pus.
One o he ou pu s o his modelling a e ocabula y g ow h cu es o
a he han ype- oken g ow h cu es, which coun he numbe o ypes
as a unc ion o he numbe o okens. A ypical shape o his cu e o
an unp oduc i e p ocess s a s wi h a ise, bu hen i “ la ens ou and
con e ges o a cons an alue, he ull ocabula y size” (2001: 168). On
he con a y, a ypical p oduc i e pa e n ea u es an in ini e ocabula y,
i.e. he cu e seems o g ow inde ini ely.
By employing he LNRE dis ibu ions, i is possible o es ima e he
numbe o ypes in la ge quan i ies o ex s, and e en he numbe o ypes
10 The LNRE heo y ( heo y o la ge numbe o a e e en s, as an independen a ea o s a is ics) has i s
o igin in 1988 when Geo gian s a is ician Es a e Khmaladze published The s a is ical analysis o a la ge
numbe o a e e en s. As his s uden Gio gi K izhinadze w o e: “The common ea u e o examples we
will p esen below is ha along wi h se e al equen e en s he e is also a e y la ge numbe o e y a e
e en s, say, wi h equency 0,1,2. The o al amoun o hese a e e en s compa ed o he numbe o ob-
se a ions ypically is no la ge bu he numbe o hese e en s among all he di e en obse ed e en s is
always e y signi ican . These a e e en s a e usually e y impo an . Fo ins ance, he numbe o wo ds
used in he book only once can be conside ed no o i al impo ance o his book, bu i is e y clea
ha hese wo ds a e absolu ely impo an because hey cons i u e hal o he au ho ’s ocabula y” (2010).
7 1Te minologija | 2021 | 28
o a wo d- o ma ion pa e n in he en i e “popula ion”, i.e. in a language
as such. A he same ime, i is possible o es ima e o p edic he numbe
o ypes in samples o he same size de i ed om la ge co po a.
As men ioned in Pa 4.1, his s a is ical p ocedu e also equi es he
quali y coun s o ypes and okens, o he wise, he calcula ion o he es i-
ma ed ocabula y size o ypes would be e oneous – ins ead o la ening
ou and con e ging o a cons an alue, he cu e could con inue o g ow.
5.1. The LNRE dis ibu ions
We decided o apply he LNRE models o iden i y11 he p oduc i i y
o e di e en ly-sized co po a, pa ly due o he ac ha he e is also an
open-sou ce ool a ailable online – Zip de eloped by e e and Ma co
Ba oni (2007). This open sou ce ool includes se e al models, om which
we used ini e Zip -Mandelb o model ( ZM) based on he Zip -Mandel-
b o law12. As he au ho s s a e, i s usage goes beyond linguis ics. Zip R ac-
cep s se e al inpu o ma s, including simple equency lis s and plain
samples, in a one- oken-pe -line o ma . We could use equency lis s om
ou p e ious esea ch ha had unde gone he manual cleaning o he 1s
s age (c . Pa 4.1), which ep esen ed ano he ad an age o his me hod.
Ou equency lis s unde wen 4 basic ope a ions (E e , Ba oni 2007):
(i) pa ame e es ima ion, whe e he pa ame e s o he LNRE model
M a e de e mined om a aining sample o size by compa ing
he expec ed equency spec um wi h he obse ed equency
spec um;
(ii) goodness-o - i e alua ion based on he co a iance ma ix o he num-
be o ypes and he numbe o ypes occu ing exac ly m imes;
(iii) in e pola ion and ex apola ion o ocabula y g ow h, using he
expec ed alues;
(i ) p edic ion o he expec ed equency spec um o a bi a y sam-
ple size.
As E e and Ba oni poin ed ou (2007), his modelling does no ocus
on he equencies o indi idual wo d ypes, “bu a he on he dis ibu ion
o such equencies (in a sample) and p obabili ies (in he popula ion)”.
11 We would like o hank o he in aluable help om ou colleague Ján Mášik who assu ed he imple-
men a ion o he ZM model o ou co pus da a.
12 Fo a comp ehensi e desc ip ion o he Zip -Mandelb o law, he eade is e e ed o, e.g. Ba oni 2009.
7 2 Jana Le ická Usage and Empi ical P oduc i i y o In e na ional
Adjec i al Su ixes in Slo ak Re isi ed
5.2. Applica ion o he LNRE dis ibu ions and e alua ion
The Zip R ool p o ided us wi h wo d equency spec a and ype- oken
cu es o each and e e y su ix in e e y (sub) co pus. Fo he lack o
space, we only p esen he g aphs showing ype- oken cu es ha we e
adjus ed o ha e he same scale on he x-axis. The y-axis di e s, as i
depends, and esul s, om he ac ual coun o ypes o a speci ic su ix
in a gi en (sub) co pus. We p opose o compa e he pai s o g aphs and
inally e alua e he p oduc i i y o su ixes ac oss ou co pus da a.
The x-axis ep esen s he numbe o okens, while he y-axis shows he
(expec ed) numbe o di e en ypes p oduced by he wo d- o ma ion
pa e n ea u ing one o he analysed su ixes.
Bo h co po a, he e e ence and scien i ic ones, show he same anking
o su ixes opped by he su ix -álny and ending wi h he su ix -ó ny.
In he e e ence co pus, he cu e o h ee su ixes -álny, -á ny, -ózny
g ows and hus indica es hei p oduc i i y, while he cu es o -i ný and,
in pa icula , -ó ny la en ou . In he scien i ic subco pus, wo su ixes
can be conside ed p oduc i e: -álny and -á ny. The cu es o he emain-
ing h ee (-ózny, -i ný, -ó ny) g ow e y slowly, bu show he endency
o la en ou a he end o he x-axis included in he g aph.
The g aphs o he medical subco pus and legal co pus show only one
p oduc i e su ix: -álny13. In he medical subco pus, bo h -á ny and -ózny
la en e y quickly (possibly, hey each hei maximum o ocabula y
size), and while -i ný and -ó ny g ow slowly, bu s eadily, hey c oss he
100- ype poin and possibly la en ou unde he 200- ype poin .
The su ix -álny in legal ex s is heading o he 350- ype poin , while in
medical ex s i is 800- ype poin (c . he scale on he y-axis). Howe e , in
legal co pus he anking sligh ly di e s – he su ix -i ný comes in he hi d
place ge ing pas he su ix -ózny, which appea s almos comple ely un-
p oduc i e wi h he cu e unning pa allel wi h x-axis. The inal su ix
-ó ny ea u es a e y slow ise and a subsequen endency o la ening.
While he g aph o he eligious co pus shows h ee p oduc i e su -
ixes -álny, -á ny and -ózny, i is no ewo hy ha all 5 analysed su ixes
seem o be p oduc i e in economic ex s. The anking o su ixes in bo h
co po a emains he same.
13 These wo g aphs miss he in e pola ion cu e o he su ix -álny. The eason o his lies in he adjus -
men and uni ica ion o he scale o he x-axis o all (sub) co po a which did no pe mi o c ea e his
ype o cu e.
7 3Te minologija | 2021 | 28
Fig. 3. Vocabula y g ow h cu es o analysed su ixes
in he medical co pus
Fig. 4. Vocabula y g ow h cu es o analysed su ixes
in he legal co pus
Fig. 1. Vocabula y g ow h cu es o analysed su ixes
in he e e ence co pus
Fig. 2. Vocabula y g ow h cu es o analysed su ixes
in he scien i ic subco pus
In he eligious co pus, bo h cu es o -i ný and -ó ny la en ou well
unde he 100- ype poin .
I we sum up he abo emen ioned indings conce ning he p oduc i -
i y o he analysed su ixes ac oss ou co pus da a, we can claim ha he
mos p oduc i e su ix is -álny, opping he anking in all he (sub) co po a.
I s p oduc i i y seems o be highes in he subco pus o scien i ic and
academic ex s and in he medical subco pus, ollowed by he e e ence
co pus ( he cu e is heading o he poin o 700 ypes), he eligious
7 4 Jana Le ická Usage and Empi ical P oduc i i y o In e na ional
Adjec i al Su ixes in Slo ak Re isi ed
co pus ( he cu e c osses he 600- ype poin ), he economic co pus ( he
cu e is heading o he 500- ype poin ) and inally he legal co pus (a
i s highes , he cu e is close o 350- ype poin ).
The su ix -á ny comes in second and ea u es p oduc i i y in 4 ou o
6 (sub) co po a. The shape o he cu es sugges s ha i is unp oduc i e
in legal and medical ex s. The highes p oduc i i y o his su ix can be
seen in he subco pus o scien i ic and academic ex s ( he cu e c osses
he 400- ype poin ), hen in he e e ence co pus ( he cu e c osses he
300- ype poin ), while in eligious and economic ex s he cu e uns
close o he 300- ype poin .
Simila ly, he su ix -ózny appea s o be p oduc i e in 4 ou o 6 co -
po a. In e es ingly enough, i seems o be unp oduc i e in he same (sub)-
co po a as he su ix -á ny – in legal and medical co po a. I is mos
p oduc i e in economic ex s whe e he cu e uns well o e he 200- ype
poin , hen in he e e ence co pus whe e he cu e c osses he same
poin , while in he emaining wo co po a he cu e ei he ouches he
same poin , bu seems o la en ou he e (p im-9.0-public-p ) o only
heads o his poin in eligious co pus.
Also, he su ix -i ný seems o be p oduc i e in 4 ou o 6 co po a. The
la ening cu e indica ing unp oduc i i y can be seen in g aphs o legal
and eligious co po a. I is mos p oduc i e in economic ex s whe e he
cu e uns well o e 100- ype poin s, while in h ee emaining (sub)-
co po a i c osses his poin and uns lowe .
Fig. 6. Vocabula y g ow h cu es o analysed su ixes
in he economic co pus
Fig. 5. Vocabula y g ow h cu es o analysed su ixes
in he eligious co pus

7 5Te minologija | 2021 | 28
The inal su ix in he anking -ó ny – is p oduc i e in 4 ou o 6 co -
po a, also. I s unp oduc i i y was iden i ied in he e e ence and eligious
co po a. I s highes p oduc i i y is in medical and economic ex s whe e
he cu e c osses he 100- ype poin , in academic and scien i ic ex s he
cu e uns close o his poin and in legal ex s i ha dly c osses he
25- ype poin .
In conclusion, i may be said ha he analysed su ixes a e mos p o-
duc i e in scien i ic and academic ex s o in he ex s o a special domain
(economy o medicine), which comes as no su p ise conside ing hei
e ymology. Thei p oduc i i y, o a he deg ee o p oduc i i y, di e s
conside ably ac oss (sub) co po a. The mos in e es ing obse a ion, pe -
haps, is he unp oduc i i y o h ee su ixes in he legal domain and he
lowes p oduc i i y o he emaining wo su ixes in he same domain. I
could indica e ha legal ex s do no a ou lexical c ea i i y in coining
new adjec i es o his kind.
Compa ed o Pa 4, he di e ence in he anking o su ixes can be
pa ially explained by he phenomenon obse ed by ácla C ček (2012)
conce ning hapax- ype a io. Acco ding o C ček’s expe imen s, his a io
ends o dec ease om i s maximal alue 1 o i s local minimum (ibid: 5).
Howe e , a e his poin , he a io s a s o inc ease again. C ček claims
i o be “some so o gene al quan i a i e p inciple o la ge collec ions o
ex s” (ibid: 14). The shape o he hapax- ype unc ion seems o be ough-
ly he same e en o ypologically di e en languages, howe e , he size
o a ex ual sample depends on he ype o he language. In o de o each
he minimal poin o hapax- ype a io, C ček s a es ha an english co -
pus should comp ise a leas 3 million okens, while a Czech one (and we
belie e a Slo ak one as well due o ypological ela edness o he wo
languages) should comp ise as many as 58 million okens. As wo subco -
po a om ou analysis a e smalle han C ček’s limi , hei hapax- ype
a io is si ua ed in he dec easing pa o he hapax- ype unc ion.
6. CONCLUSION
In he sequel o ou esea ch ocusing on he p oduc i i y o selec ed
su ixes in di e en co po a and domains, we managed o answe bo h
ques ions p esen ed in Pa 4.
1. We we e able o p o e ha low- equency lemmas ex ac ed om a
co pus also include po en ial neologisms, and, hough hei dis ibu ion
7 6 Jana Le ická Usage and Empi ical P oduc i i y o In e na ional
Adjec i al Su ixes in Slo ak Re isi ed
may a y ac oss co pus da a, hey can al e he p oduc i i y anking based
on hapax/ ype o mula.
2. We iden i ied and applied s a is ical modelling o he e alua ion o
p oduc i i y ac oss co po a – LN e dis ibu ions. By eeding he open-
sou ce ool wi h ou co pus da a we ecei ed a isual modelling o p o-
duc i i y ha enabled us o s a e ha he p oduc i i y o he analysed
su ixes di e s no only when gene al and scien i ic ex s a e compa ed,
bu also be ween di e en specialised domains. We belie e ha he mod-
elling clea ly shows he eal po en ial o each and e e y analysed su ix
o p oduce new ypes in a espec i e domain.
Howe e , he esul s and anking om Pa 4 and 5 o his pape a e
a he mu ually incompa able because he i s one is based di ec ly on
aw da a – he coun o ypes and hapaxes o low- equency ypes in a
speci ic (sub) co pus, while he second one on he expec ed equency
spec a o e e y (sub) co pus. In sho , while he i s one ep esen s ac-
ual da a, he second one seeks o es ima e he p oduc i i y o an elemen
in he speci ic language as such. The e o e, o la ge co po a o medical
and legal ex s, we should iden i y mo e hapaxes, which migh esul in a
di e en su ix anking. Ou assump ions could e en ually be e i ied in
u u e esea ch by analysing mo e ex ensi e co po a comp ising a ep e-
sen a i e sample o ele an ex ypes o speci ic domains because ex
gen es can also play a ole in mo phological p oduc i i y o an elemen .
Gi en ha e en he bigges co pus would no comp ise all he wo ds,
we belie e ha i is mo e easonable o e alua e he mo phological p o-
duc i i y o any elemen by means o mo e sophis ica ed me hods han
he aw coun o ypes o okens.
In conclusion, i could be said ha we a e able o iden i y he pas
p oduc i i y in a wo d- o ma ion pa e n o elemen . We can also p o ide
di e en es ima es o i s u u e exploi a ion, i.e., Baayen’s po en ial p o-
duc i i y. Fu he mo e, e en i we comple e he s a is ical measu es wi h
a quali a i e analysis, we should ake in o conside a ion one mo e ac : as
Fe nández-Domínguez igh ly a gues, p obabilis ic p edic ions o mo -
phological p oduc i i y should be pe cei ed in he con ex o ex alin-
guis ic ac o s and, in pa icula , hose in ol ing he naming needs o a
speech communi y. “I no naming need exis s, no p oduc i e wo d- o -
ma ion can ake place” (Fe nández-Domínguez 2013: 438).
7 7Te minologija | 2021 | 28
ACKNOWlEDgEMENTS
The pape has been w i en wi hin he Slo ak Na ional Co pus p ojec suppo ed
by he Slo ak Academy o Sciences, Minis y o educa ion, Science, esea ch and
Spo o he Slo ak epublic, Minis y o Cul u e o he Slo ak epublic and he
Ľudo í Š ú Ins i u e o Linguis ics, Slo ak Academy o Sciences.
lIST OF SOURCES
Slo enský ná odný ko pus. Ko pus p im-7.0- k. B a isla a: Jazyko edný ús a Ľ. Š ú a SA 2018. Acces-
sible a : h ps://ko pus.sk.
Slo enský ná odný ko pus. Ko pus p im-9.0-public-p . B a isla a: Jazyko edný ús a Ľ. Š ú a SA 2020.
Accessible a : h ps://ko pus.sk.
Slo enský ná odný ko pus. Ko pus p im-9.0-juls-all. B a isla a: Jazyko edný ús a Ľ. Š ú a SA 2020. Ac-
cessible a : h ps://ko pus.sk.
Slo enský ná odný ko pus. Ko pus legal-1.1. B a isla a: Jazyko edný ús a Ľ. Š ú a SA 2013. Accessible
a : h ps://ko pus.sk.
Slo enský ná odný ko pus. Ko pus bl -2.0. B a isla a: Jazyko edný ús a Ľ. Š ú a SA 2014. Accessible a :
h ps://ko pus.sk.
Slo enský ná odný ko pus. Ko pus ecn-2.0-public. B a isla a: Jazyko edný ús a Ľ. Š ú a SA 2016. Acces-
sible a : h ps://ko pus.sk.
Slo níko ý po ál Jazyko edného ús a u Ľ. Š ú a SA h ps://slo nik.juls.sa ba.sk/.
The Zip R package. A ailable a : h p://zip . - o ge. -p ojec .o g/.
REFERENCES
Assadi Houssem, Bou igaul Didie 1995: Classi ica ion d’adjec i s ex ai s d’un co pus pou l’aide à la mo-
délisa ion de connaissances. – T oisièmes jou nées in e na ionales d’analyse des données ex uelles, 313–320.
Baayen . Ha ald, Liebe ochelle 1991: P oduc i i y and english de i a ion: a co pus-based s udy. – Lin-
guis ics 29, 801–843.
Baayen . Ha ald 1992: quan i a i e aspec s o mo phological p oduc i i y. – Yea book o Mo phology 1991,
eds. g. e. Booij, J. an Ma le, Do d ech : Kluwe Academic Publishe s, 109–149.
Baayen . Ha ald 1993: On equency, anspa ency, and p oduc i i y. – Yea book o Mo phology 1992,
eds. g. e. Booij, J. an Ma le, Do d ech : Kluwe Academic Publishe s, 181–208.
Baayen R. Ha ald 1994: P oduc i i y in p oduc ion. – Language and Cogni i e P ocesses 9, 447–469.
Baayen R. Ha ald, Renou An oine e 1996: Ch onicling The Times: P oduc i e Lexical Inno a ions in an
english Newspape . – Language 72, 69–96.
Baayen . Ha ald, Liebe ochelle 1997: Wo d equency dis ibu ions and lexical seman ics. – Compu e s
and he Humani ies 30, 281–291.
Baayen . Ha ald 2009: Co pus linguis ics in mo phology: mo phological p oduc i i y. – Co pus Linguis ics.
An in e na ional handbook, eds. A. Lüdeling, M. Ky o, Be lin, Mou on De g uy e , 900–919.
Bake Paul, Ha die And ew, Mcene y Tony 2006: A Glossa y o Co pus Linguis ics, Edinbu gh: Edinbu gh
Uni e si y P ess L d. A ailable a : h ps://pd s.seman icschola .o g/856a/9640311005 8a97ab98976c9209
aa12120a.pd .
Ba oni Ma co 2009: Dis ibu ions in ex . – Co pus Linguis ics. An in e na ional handbook, eds. A. Lüdeling,
M. Ky o, Be lin, Mou on De g uy e , 803–822. A ailable a : h ps://home.sslmi .unibo.i /~ba oni/pub-
lica ions/hsk_39_dis _ e 2.pd .
Ca iè e Isabelle 2008: Médi e m: encodage des adjec i s médicaux. – Co pus e dicionnai es de langues de spéciali-
és, eds. F. Manize, P. Du y, N. A lin, C. ougemon , g enoble: P esses uni e si ai es de g enoble, 175–196.
C ček ácla 2012: How La ge is he Co e o Language. – Co pus Linguis ics 2011 [online], Bi mingham.
A ailable a : h ps://www.bi mingham.ac.uk/Documen s/college-a slaw/co pus/con e ence-a -
chi es/2011/Pape -145.pd .
Daille Béa ice 2001: quali a i e e minology ex ac ion: Iden i ying ela ional adjec i es. – Recen ad ances
in compu a ional e minology, eds. D. Bou igaul , C. Jacquemin, M.-C. L’Homme, Ams e dam: John Ben-
jamins Publishing Co, 149–166.
7 8 Jana Le ická Usage and Empi ical P oduc i i y o In e na ional
Adjec i al Su ixes in Slo ak Re isi ed
Dokulil Miloš 1962: Teo ie oření slo , P aha: Naklada els í ČSA .
E e S e an, Lüdeling Anke 1991: Cons aining psycholinguis ic models o mo phological p ocessing and
ep esen a ion: he ole o p oduc i i y. – Yea book o mo phology 1991, eds. g. Booij, J. an Ma le, Do -
d ech : Kluwe Academic Publishe s, 165–183.
e e S e an, Ba oni Ma co 2007: The zip R lib a y: Wo ds and o he a e e en s in R. P esen a ion a useR!
2006: The Second R Use Con e ence, Vienna, Aus ia. A ailable a : h ps://zip . - o ge. -p ojec .o g.
e e S e an 2004: A simple LN e model o andom cha ac e sequences. – P oceedings o JADT 2004,
411–422.
Fe nández-Domínguez Jesús 2013: Mo phological P oduc i i y Measu emen : explo ing quali a i e e sus
Quan i a i e App oaches. – English S udies 94, 4, 422–447.
ga abík ado an, giani so á Lucia, Ho ák Alexande , Šimko á Má ia 2004: Tokenizácia, lema izácia a mo o-
logická ano ácia Slo enského ná odného ko pusu: in e nal documen o manual mo phological anno a ion.
Unpublished. A ailable a : h ps://ko pus.sk/a achmen s/publica ions/2004-ga abik-giani so a-ho ak-sim-
ko a- okenizacia.pd .
Ho ecký Ján 1956: Základy slo enskej e minológie, B a isla a: VEDA.
Ho ecký Ján 1999: In e nacionalizácia a eu opeizácia slo enčiny. – In e nacionalizácia súčasných slo an-
ských jazykoch: za a p o i, ed. J. Bosák, B a isla a: eda, 80–82.
Hulse Vic o ia 2011: P oduc i i y in mo phological nega ion: a co pus based app oach, The uni e si y o Man-
ches e . A ailable a : h ps://www. esea ch.manches e .ac.uk/po al/en/ heses/p oduc i i y-in-mo pholog-
ical-nega ion-a-co pusbased-app oach(266d2241-a266-4b99-8 ab-e19571381d8 ).h ml.
K izhinadze Gio gi 2010: La ge numbe o a e e en s: Di e si y analysis in mul iple choice ques ionnai es and
ela ed opics, Vic o ia Uni e si y o Welling on doc o hesis. A ailable a h ps://co e.ac.uk/download/
pd /41336663.pd
Le ická Jana 2021: Usage and empi ical p oduc i i y o in e na ional adjec i al su ixes in Slo ak based on
gene al and specialised co po a. – Jazyko edný časopis 72, 2 (in p in ).
L’Homme Ma ie-Claude 2002: Fonc ions lexicales pou ep ésen e les ela ions en e e mes. – T ai emen
au oma ique des langues 43, 1, 19–41.
L’Homme Ma ie-Claude 2003: Adjec i s dé i és séman iques dans la s uc u a ion des e minologies. –
Jou nées d’é ude Te minologie, On ologie e ep ésen a ion des connaissance, Lyon.
Maniez F ançois 2002: Dis ingue les e mes des colloca ions: é udes su co pus du pa on < Adjec i –
Nom > en anglais médical. – Ac es du colloque TALN de Nancy 1, 24–27 juin 2002, 345–350.
an Ma le Jaap 1992: The ela ionship be ween Mo phological P oduc i i y and F equency: A Commen on
Baayen’s Pe o mance-O ien a ed Concep ion o Mo phological P oduc i i y. – Yea book o Mo phology
1991, eds. g. Booij, J. an Ma le, Do d ech : Kluwe Academic Publishe s, 151–163.
Nábělko á Mi a 1996: a ian nosť ako p ok dynamiky adjek í nej pa adigme. – Slo enská eč 61, 257–266.
Nacca a o Chia a 2016: A co pus-based quan i a i e app oach o he s udy o mo phological p oduc i i y in
diach ony: The case o samo-compounds in ussian. – A Blend o MaLT Hanna Ch is , eds. D. Kle no šak,
L. Sönning, . We ne , Bambe g: uni e si y o Bambe g P ess, 133–153.
No mand Syl ie, Bou igaul Didie 2001: Analysing adjec i es used in a his opa hology co pus wi h NLP
ools. – Te minology 7, 2, 155–164.
Plag Ingo, Dal on-Pu e Ch is iane, Baayen Ha ald 1999: Mo phological p oduc i i y ac oss speech and
w i ing. – English Language and Linguis ics 3, 2, 209–228.
Säily Tanja 2018: Change o a ia ion? P oduc i i y o he su ixes -ness and -i y. – Pa e ns o Change in
18 h Cen u y English. A Sociolinguis ic App oach, eds. T. Ne alainen, M. Palande -Collin, T. Säily, Am-
s e dam: John Benjamins, 197–218.
Še číko á Magda 2014: Zjišťo ání slo o o né p oduk i i y z ko puso ých da : přípony od ozující náz y
las nos í. – Naše eč 97, 228–240.
Š ícha F an išek 2012: Jak epoše elek onických ko pusů následo a Miloše Dokulila (Miloši Dokulilo i ke
s ému ý očí na ození). – Jazyko ědné ak uali y 49, 95–107.
Š ícha F an išek 2002: K Dokulilo u pojmu slo o o né p oduk i i y (z hlediska ko puso é analýzy). –
Češ ina doma a e s ě ě 4, 302–310.
Š ícha F an išek 2007: Ko puso é s a is iky a slo o o ná p oduk i i a. – G amma & Co po a/G ama ika a
ko pus 2005, eds. F. Š ícha, J. Šimandl, P aha: Academia, 250–257.
Š ícha F an išek 2009: Slo o o ná p oduk i i a a g ama ičnos : g adační exp esi ní adjek i a s p e ixy
p a-, pře- a ele- současné psané češ ině. – Esla ís ica Complu ense 9, 145–170.