Documente Academic
Documente Profesional
Documente Cultură
Loanwords in Romanian*
Kim Schulte
!
The subdatabase of the World Loanword Database that accompanies this chapter is available online
at http://wold.livingsources.org. It is a separate electronic publication that should be cited as:
Schulte, Kim. 2009. Romanian vocabulary. In Haspelmath, Martin & Tadmor, Uri (eds.) World
Loanword Database. Munich: Max Planck Digital Library, 2137 entries.
<http://wold.livingsources.org/vocabulary/8>
Romanian can be subdivided into two major dialect groups, the Muntenian-based
dialects spoken in the south, and the Moldavian-based ones spoken in the north of
Romania and the Republic of Moldova. The official name of the national language
of the Republic of Moldova is Moldovan or Moldavian, but linguistically speaking it
is very similar to the neighboring dialects of north-eastern Romania. In general,
Romanian has comparatively little dialectal variation, but regional differences can
nevertheless be observed and are the basis for a distinction between dialects such as
Moldavian, Transylvanian, or that of the Banat region. An important distinguishing
feature between regional varieties is their lexicon, particularly lexical loans; unsur-
prisingly, those donor languages spoken in the immediate vicinity tend to be the
source of a comparatively larger proportion of loanwords in the respective regional
varieties.
In order to provide loanword data for Romanian as a whole, the lexicon used for
the subdatabase is not based on any specific regional dialect, but on what is consid-
ered to be part of the language according to the Romanian Academys dictionary
(Coteanu et al. 1998). As a result, a number of the loanwords included are most
commonly used in particular regions; in some cases this results in the incorporation
of several synonyms borrowed from different source languages.
The historic foundations for the emergence of Romanian were laid when the
Dacians, inhabitants of an area broadly coinciding with modern-day Romania, were
defeated by the Romans under Emperor Trajan between 101 and 106 CE, leading
to the foundation of the Roman province of Dacia. This was followed by a period
of intense colonization and Romanization, during which a regional variety of Popu-
lar Latin established itself as the local language. The contact with the rest of the
Roman Empire was relatively short-lived, as the invading Goths forced Rome to
pull out of Dacia after less than 170 years, around 271 CE. Despite the compara-
tively short duration of direct contact with the rest of the Roman Empire, language
shift from the Thraco-Dacian substrate to Latin must have been sufficiently exten-
sive for a Latin-based language that we might call proto-Romanian to completely
replace the substrate language(s), though this may have been a gradual and pro-
longed process (see 3.1 below).
Subsequently, various peoples invaded the area, generally moving in from the
northeast and east. Whilst some invading tribes, e.g. the Huns, left few cultural and
linguistic traces, other populations settled amongst the early Romanian speakers,
th th
notably Magyars (from the 9 century) and Slavs in several waves of migration (6 -
th
11 century), providing ideal conditions for long-term linguistic contact. There is
an ongoing debate as to whether ethnic Romanians have been living in areas north
of the Danube, particularly in Transylvania, continuously since Roman times, or
whether they were pushed back by a large Hungarian population, eventually return-
ing to those areas at a later stage. As this debate is primarily politically motivated
and linked to territorial claims, it will not be entered into here; in any case, the
linguistic evidence suggests a considerable degree of cultural contact, typical of a
situation of cohabitation over an extended time period.
Other linguistically relevant historical events include the arrival of German set-
th th
tlers in Transylvania in the 12 and 13 centuries, encouraged by the Hungarian
th
rulers, and the imposition of Ottoman suzerainty from the 16 century, bringing
the population into increased cultural, administrative and trade-based contact with
other areas of the Ottoman empire, particularly modern-day Turkey, Bulgaria, and
Greece.
2. Sources of data
The source of the lexical data, i.e. the Romanian words corresponding to the
Loanword Typology meanings, was either the authors personal knowledge or stan-
dard bilingual dictionaries (Isb"!escu 1995; Savin et al. 1997; Levi#chi & Banta!
1992), complemented by the Romanian Academys monolingual dictionary
(Coteanu et al. 1998) and a dictionary of synonyms (Seche & Seche 1997); the lat-
ter were used to identify any existing synonyms and to determine the degree of
semantic overlap between near and partial synonyms.
The two main sources of the etymological information that appears in the data-
base are (a) the Romanian Academys Dic!ionarul explicativ al limbii romne
(Coteanu et al. 1998), which provides the source language and etymon, where
known, for each entry, but does not supply any additional etymological explanation
or discussion, and (b) Cior"nescus (1966) etymological dictionary of Romanian,
which contains very detailed etymologies but has a limited number of entries.
Where neither of these default sources provided a fully satisfactory etymology, it
was either complemented with suggestions by time-honored Romanian philologists
(Pu!cariu 1943 [1997]; Philippide 1894; Ha!deu 1877, 1879, 1883), or more specific
studies dealing specifically with the etymology of loanwords from individual source
languages were consulted. Among these, Wendt (1960) examines loans from Turk-
ish, Miklosich (1860, 186265) investigates the incorporation of Slavic elements
into Romanian, Conev (1921) looks at contact between Bulgarian and Romanian,
Murnu (1894) and Diculescu (192426) investigate Greek elements in Romanian,
whilst Cihac (1879) and McClure (1976) examine the loans from various source
languages. Information regarding the exact word form of the source word was fre-
quently taken from dictionaries of the respective languages, e.g. Newmark (1998)
and Fiedler & Klosi (1997) for Albanian, Gruji$ (1998) for Serbian, and Steuerwald
(1972) for Turkish. Information regarding the earliest known source word was
obtained from etymological dictionaries of various languages, e.g. Corominas (1961)
for loanwords shared with Spanish and Grebe (1963) for loanwords with German
cognates.
3. Contact situations
For the present analysis of loanwords in Romanian, the focus lies on words bor-
rowed into the language after Latin began to be used in the area where Romanian is
spoken today. Whilst it is neither possible nor sensible to define an exact point in
time at which Latin became Romanian, it can be ruled out that a Latin-based
Romanian language existed before Latin began to be used in the territory. Thus,
loanwords in Latin such as camisia shirt, alb from Germanic hemidi mantel, shirt,
which entered Latin via Celtic and was passed on, like any native item, to its Ro-
mance daughter languages (Spanish camisa shirt, French chemise shirt, Romanian
c"ma#" shirt), are certainly borrowed, but into Latin, not into Romanian. As the
aim of this chapter is to examine the impact of borrowing on the lexical structure
of Romanian and to compare and contrast these developments with other languages
(including Romance sister languages of Romanian), the following survey examining
contact situations that have left their traces in the Romanian lexicon will begin
around the time when a specifically Dacian or Romanian regional variety of Popular
Latin began to develop, namely in the second century CE, after the Roman con-
quest of the area.
Little is known about the Thraco-Dacian substrate spoken in the area before the
shift to Latin, but it is generally assumed that it was an Indo-European language
closely related to Albanian, perhaps even the direct ancestor of modern Albanian
(du Nay 1996: 72). Whilst this Thraco-Dacian substrate disappeared with the
adoption of Latin, it must be assumed that contact with closely related languages,
perhaps varieties of Albanian, continued for several centuries. In addition to peas-
ants in remote areas, who were not immediately affected by the Roman occupation
and probably took longer to shift to Latin, a certain degree of population move-
ment and mixture between Latinized and non-Latinized areas must be assumed,
particularly due to the fact that semi-nomadic herdsmen roamed large areas of the
Balkan Peninsula, thereby acting as a continuous source of contact with the
Thraco-Dacian/Albanian language(s). In most cases, our lack of precise knowledge
of these languages makes it impossible to determine whether Romanian words with
cognate counterparts in Albanian were borrowed directly from the substrate lan-
guages, or from Albanian at a later stage (Rosetti 19381941 [1978: 223]).
The nature of the contact between Latin and the local substrate language can be
assumed to have proceeded along similar lines as in many other areas that were
incorporated into the Roman Empire. After conquering and occupying the territory
militarily, Roman administrative structures were implemented and former Roman
soldiers from across the Empire were given land on which to settle. For the existing
inhabitants of the area, who received the status of Roman citizens, it was useful or
even necessary to learn and speak Latin to participate and be successful in this new
society; there was little resistance to adopting Latin, as it had more prestige than
the substrate language and was associated with wealth and progress. This resulted
in a rapid language shift to Latin, despite the relatively brief period of Roman rule
(106 CE to 271 CE); the number of lexical and morpho-syntactic elements retained
from the substrate (i.e. borrowed into the regional variety of Latin) is comparatively
1
small , despite some ongoing contact with languages closely related to the original
substrate, such as Albanian.
The relationship between the Latin superstrate and the local substrate languages
was initially one defined by the political and cultural dominance of the Romans.
After the Romans withdrawal from the province, however, it can be assumed that
the contact situation gradually changed to one of cohabitation, in which speakers of
early Romanian and speakers of Thraco-Dacian/Albanian lived in close vicinity of
each other and communicated on a regular basis about everyday matters regarding
their pastoral activity and the natural environment.
1
A number of morpho-syntactic elements are shared by several languages belonging to the Balkan
Sprachbund, some of which may be rooted in structures that were present in the common substrate
language(s) (see 6).
In addition to the comparatively early South Slavic influence, there was more local-
ized contact between Romanian and individual Slavic languages at later dates. There
is evidence of contact between Romanian and Ukrainian that must have taken place
th
after the 12 century (as shown by the fact that borrowing took place after a
th
Ukrainian [h]>[g] change, dated around the 12 century, cf. Mih"il" 1973: 46) in
th
the north, of contact with Serbian since the 15 century in the east, as well as con-
tinuing contact with Bulgarian in the south. These regionally limited contact
situations were characterized by interaction, in most domains of everyday life, be-
tween the Slavic and the Romanian populations in the respective areas.
Contact between Romanian and (Byzantine) Greek was both direct and indirect.
th
From before the Byzantine period until approximately the 10 century, Balkan
Romance, as well as South Slavic, were spoken in an area that bordered on northern
Greece; in the south of this area, the presence of a considerable Greek population
led to a trilingual contact situation. Cohabitation and everyday interaction between
all three population groups were common. Even after Daco-Romanian was physi-
cally separated from the Greek-speaking area, contact between Romanian and Greek
continued, especially in the areas of trade and commerce. To what extent this con-
tact was mediated by the Slavic population that separated the two cannot be
precisely determined; the presence of numerous Greek loans in Romanian, Bulgar-
ian and other Balkan languages shows that these words were widely borrowed
throughout the region.
th
From the 15 century onwards, contact between Greek and Romanian speakers,
primarily through commercial activity, continued within the expanding Ottoman
Empire.
largely excluded from official matters; contact with Hungarian thus remained com-
paratively limited.
Having been granted special privileges, German settlers founded towns and villages
th th th th
in Transylvania in the 12 and 13 centuries. In the 17 and 18 centuries, more
German settlements were founded in the Banat area in the east of Romania. Con-
tact with the Romanian population was similar to that between Romanians and
Hungarians, as the German settlers preserved their separate cultural and linguistic
identity. Contact was generally limited to commercial interaction.
th
A separate contact situation, beginning in the second half of the 19 century,
arose due to an increasing orientation towards Western European culture and life-
style. Though the primary cultural model was France, Germany also served as a
model. The upper classes travelled to Germany and visited German universities,
German literature was read in intellectual circles, and certain novel products and
concepts came to Romania from or via Germany.
3.6.1. Trade
th
Turkish traders settled along the Black Sea coast from the late 15 century on-
wards. Being a relatively small group of the population, the overall impact of this
contact was largely limited to commercial transactions.
The Ottoman rule in the Balkans brought about a contact situation in which
Turkish acquired considerable importance as a language used in the military and
administrative domains. Whilst Romanian-speaking areas retained a certain degree
th th
of independence under Ottoman suzerainty between the 16 and 18 centuries,
many Turkish products and cultural practices found their way into Romania. Fur-
thermore, Romanians were recruited to fight in the Ottoman armies, which led to a
high degree of contact in the military domain.
Like in many European cultures, Latin, and by extension Italian, was viewed as an
th
educated linguistic model. In Romanian, this trend surfaces as early as the 17
century, exemplified by the borrowing of Italian popolo > Romanian popor people,
population, as an alternative to neam, borrowed from Hungarian.
The cultural importance of Latin, Italian, to some extent German, but most of
th th
all French, especially during the second half of the 19 and the first half of the 20
Of the words contained in the subdatabase for Romanian, about 42% are loan-
words. Whilst this implies that the majority of the Romanian lexicon is inherited
from Latin, it also shows that Romanian has incorporated an exceptionally large
amount of lexical material from other languages. Table 1 shows the distribution of
loanwords by donor language and semantic field.
Perhaps unsurprisingly, the largest number of loanwords is found in the seman-
tic field Modern world, where 70.5% of the words are borrowed. About half of these
loanwords come from French, either fully or partially. (For a discussion of the con-
cept of partial borrowing from French, see 4.2 below.) Due to the strong cultural
th th
orientation towards France in the 19 and 20 centuries (see 3.7), the reason for
this intense borrowing from French can be linked, at least in part, to the fact that
many of the new inventions and concepts belonging to this semantic field were
introduced to Romanian speakers through France and the French language. This is
also the background for the only borrowed item in the category Miscellaneous
function words: a deveni to become, an alternative to the synonymous periphrastic a
se face (literally to make oneself) was borrowed from French devenir as a term
linked primarily to modern philosophy, which Romanians came into contact with
through French.
On the other hand, loans from the Slavic languages are hardly relevant in the
category Modern world, despite accounting for a significant proportion of the bor-
rowed lexical stock in general. Items in this category, it may be assumed, were not
normally introduced to Romanian speakers via the populations of neighboring
countries during the period in which most of the Modern world meanings emerged
and gained significance. In this semantic field, the largest amount of loans from a
Slavic language come from Russian, largely due to Russia serving as a model in the
area of administration. Thus po#ta post, mail and poli!ie police are borrowed from
Russian.
Table 1:
Loanwords in Romanian by donor language and semantic field (percentages)
Ukrainian/Russian
Total loanwords
Non-loanwords
Hungarian
Bulgarian
Albanian
Turkish
German
Serbian
French
Others
Italian
Greek
Slavic
Latin
1 The physical 17.5 12.3 6.3 0.4 1.2 2.4 0.8 - - 0.8 4.8 0.8 - 47.2 52.8
world
2 Kinship 6.8 5.3 4.5 0.5 0.5 1.1 1.1 - - 0.3 3.2 - - 23.1 76.9
3 Animals 16.9 4.7 4.4 1.7 2.4 2.0 2.0 1.4 1.0 1.7 2.7 1.4 0.7 42.9 57.1
4 The body 17.8 4.4 3.5 2.7 1.9 0.8 1.9 2.3 1.2 1.2 1.2 - 0.4 39.2 60.8
5 Food and drink 6.4 4.5 7.3 0.9 - 4.5 2.7 - 0.9 1.8 - 1.8 - 30.9 69.1
6 Clothing and 19.0 5.6 7.8 1.1 3.4 4.5 6.1 4.5 3.4 2.8 1.1 2.2 1.1 62.6 37.4
grooming
7 The house 4.3 10.0 8.2 - 1.4 10.0 5.7 4.3 1.4 5.4 1.4 2.9 1.4 56.4 43.6
8 Agriculture and 14.6 6.6 8.0 1.4 - 4.7 - 1.9 2.8 6.1 2.8 1.8 0.9 51.9 48.1
vegetation
9 Basic actions 12.6 13.9 4.1 0.9 1.3 1.7 0.4 3.5 1.7 1.1 - 3.5 - 44.6 55.4
and technology
10 Motion 9.6 11.0 3.5 2.2 3.5 0.9 1.8 0.9 0.9 0.9 - - - 35.1 64.9
11 Possession 10.6 12.1 4.2 4.5 7.6 - 1.5 - 1.5 2.7 - - - 44.7 55.3
12 Spatial relations 12.2 2.6 1.3 3.9 - - - - 3.0 0.4 - 0.9 - 24.3 75.7
13 Quantity 2.1 9.4 - 4.2 - 2.1 - - - - - - - 17.7 82.3
14 Time 3.9 12.3 1.9 3.9 1.3 - - 1.3 - 0.6 - - - 25.3 74.7
15 Sense perception 2.6 5.2 0.9 0.9 - 1.7 - - - 2.6 - - - 13.9 86.1
16 Emotions and 16.8 14.7 2.1 3.7 5.3 - 2.1 - 0.5 - - 2.1 - 47.4 52.6
values
17 Cognition 16.8 6.9 1.7 8.1 2.3 - 1.2 3.5 1.7 1.7 - - - 43.9 56.1
18 Speech and 9.8 16.5 0.8 7.5 2.3 - - 4.5 - 0.8 - 1.5 1.5 45.1 54.9
language
19 Social and po- 19.4 17.7 4.0 11.3 1.6 3.2 - 6.5 - 0.8 1.6 - - 66.1 33.9
litical relations
20 Warfare and 16.0 9.6 3.2 4.8 4.8 3.2 - 1.6 1.6 1.6 - 1.6 1.6 49.6 50.4
hunting
21 Law 24.1 5.1 - 7.6 2.5 - 2.5 2.5 - - - - - 44.3 55.7
22 Religion and 10.2 22.7 2.3 13.6 - 2.3 4.5 - 3.4 - - - - 59.1 40.9
belief
23 Modern world 36.3 0.7 1.5 4.4 2.2 - 5.8 - 15.3 - - 4.4 - 70.5 29.5
24 Miscellaneous 6.5 - - - - - - - - - - - - 6.5 93.5
function words
13.7 8.4 3.9 3.2 1.9 1.9 1.7 1.6 1.6 1.5 1.0 1.1 0.3 41.8 58.2
In the category Religion and beliefs, there is also a large proportion of loanwords,
almost 60%. Half of these are borrowed from early South Slavic, Bulgarian or
Greek, i.e. the languages through which Romania was in touch with the Orthodox
Church. The use of Old Church Slavonic as the language of religion for many cen-
turies explains the large number of Slavic loans; it is likely that words belonging to
other semantic fields also entered the language via this religious use of Slavic, for
instance dragoste love and prieten friend.
Another semantic field with an exceptionally large proportion of loans, 66.1%,
is Social and political relations. This is not entirely unexpected, as new social and
political structures are often influenced by, or imported from, populations with a
different socio-political system when a contact situation arises. Even subtle differ-
ences between the old concepts and the newly imported or adapted ones are likely
to be reflected lexically, as the affected speakers are acutely aware of the differences
affecting their daily lives. Many loanwords in the category Social and political
relations, around a third of the loanwords in this category, come from Slavic lan-
guages, which had the most profound impact on Romanian society over the
centuries. These words include rather fundamental concepts such as a porunci to
order, to command as well as st"pn master and rob slave, which suggest that
contact with the Slavs brought changes to this aspect of society with it; the influ-
ence of the use of Old Church Slavonic in religious contexts is also likely to have
contributed to the adoption of these words. Another case showing how differences
that came with a new administrative and political system are reflected by lexical
replacement is grani!" border, frontier, also borrowed from Slavic, which may be
assumed to have come into use with a new type of frontier established under the
influence of a Slavic administrative system.
An example of a distinction between similar meanings being made by means of
adopting a loanword is the near-synonymous word pair for village, in which c"tun,
cognate with Albanian katund, is probably a loan from the pre-Latin substrate and
refers to a hamlet without any formal administrative structure of its own, a type of
settlement that may be assumed to have existed before Roman occupation. The
inherited Romance word for village, sat, derived from Latin fossatum, on the other
hand, refers to a somewhat larger village which, according to the meaning of
fossatum, was originally typically fortified in some way. Furthermore, the much later
borrowing of ora# town, city from Hungarian vros suggests that urbanization in
the present-day sense came to Romanian primarily through contact with Hungarian
speakers.
Apart from Miscellaneous function words, the semantic fields with the lowest per-
centage of loanwords are Sense perception (13.9%) and Quantity (17.7%). In both of
these categories, the majority of borrowed items come from Slavic; the proportion
of loans from Slavic in almost all semantic fields will be further discussed in 4.2
below.
All in all, it is significant to observe that borrowing into Romanian has occurred
across the entire lexicon; the average of close to 42% across all semantic fields is
not distorted by exceptionally large numbers of loans in particular semantic areas.
This is confirmed by the fact that the median percentage of loanwords across all
categories is approximately 45%, indicating that a high proportion of borrowed
items is found across most of the semantic fields distinguished in the loanword
database.
Sorted by semantic word class, Romanian loanwords conform to the common
pattern that nouns appear to be most easily borrowed, as shown by Table 2. Just
over 50% of the nouns in the subdatabase are loans, whilst verbs and adjectives have
an almost equal loanword quota of 32%. As adjectives and adverbs are generally not
morphologically distinguished in Romanian, the proportion of borrowed adverbs
must be assumed to be approximately equivalent to that of borrowed adjectives.
Table 2:
Loanwords in Romanian by donor language and semantic word class (percentages)
Ukrainian/ Russian
Total loanwords
Non-loanwords
Hungarian
Bulgarian
Albanian
Turkish
German
Serbian
French
Others
Italian
Greek
Slavic
Latin
Nouns 16.1 8.2 5.1 2.9 2.4 2.9 2.6 2.0 2.5 1.9 1.7 1.5 0.5 50.2 49.8
Verbs 10.5 11.2 2.1 4.3 1.3 - 0.4 1.4 - 0.5 - 0.2 - 32.1 67.9
Adjectives 13.6 7.1 2.1 3.0 1.3 1.0 - 0.5 0.8 1.6 - 1.0 - 32.0 68.0
Adverbs - 20.0 - - - - - - - - - - - 20.0 80.0
Function words 1.3 0.8 0.4 2.1 - - - 0.8 - 0.4 - - - 5.9 94.1
all words 13.7 8.4 3.9 3.2 1.9 1.9 1.7 1.6 1.6 1.5 1.0 1.1 0.3 41.8 58.2
The largest number of loanwords comes from French. About 12% of the
Romanian words in the database are unambiguously borrowed from French, and
this number rises to about 16% if all loanwords that have French as a partial or
possible source are added. A word can be considered to be partially borrowed from
French if its form in Romanian does not allow us to determine unambiguously
whether it is borrowed from French or from one of the other languages that served
as cultural models at the same time. An example of a word borrowed from French
and Latin is Romanian conspira!ie plot, conspiracy, which is morphologically inte-
grated in such a way that it is not evident whether it is borrowed from Latin
conspiratio or French conspiration; both of these would change to conspira!ie accord-
ing to the normal rules of loanword integration. In such cases, it is quite possible
that the Romanians who had knowledge of both the possible source languages be-
gan to use the word in Romanian without consciously deciding from which of the
two languages they had borrowed the word. Similarly, Romanian banc" bank
might be borrowed from Italian banca and French banque, and Romanian ren rein-
deer might be borrowed from German Ren and French renne.
If we further add learned Latin and Italian loanwords to the French ones, then
the total proportion of items borrowed from languages serving as cultural models is
about 20%, i.e. one in five.2 Such a high proportion of educated loans requires
some explanation. Whilst it is normal for a speech community to borrow terms for
newly introduced objects or concepts, Romanian has gone far beyond this level,
borrowing heavily from French to create synonyms of words already present in the
language. Thus Romanian surs" spring from French source is synonymous with
izvor, borrowed from Slavic izvor$, and with fntn" (now archaic in this meaning),
the word inherited from Latin. Similarly, Romanian litoral coast has been bor-
rowed from French littoral despite the existence of the synonyms coast" and !rm,
inherited from Latin costa and termen, as well as Romanian mal from Alba-
nian/substrate malj. In such cases, it is relevant to determine whether the newly
borrowed item is merely a marginal, rarely used or stylistically restricted alternative,
or whether it is frequently used and in real competition with the synonyms that it
is predated by. However, the reality is not always quite as black and white. In the
case of coast, for example, mal is used 50% more than litoral, but litoral still occurs
twice as frequently as coast"; !rm is only found in just over 1% of all cases. This
means that the most recently added synonym, the loanword from French, has es-
tablished itself as a serious competitor.
In some cases, two cognate words have been borrowed into Romanian from
more than one of these languages. For example, jaluzie jealousy, borrowed from
French jalousie, has been borrowed in addition to the more commonly used gelozie,
from Italian gelosia; there is also a synonym borrowed from Greek, zulie. All of
these are winning the competition with the inherited terms temere or temut, which
covers a somewhat wider emotional area including fear as well as jealousy.
Loans from French, Latin, Italian, and to some extent from German are found
in all semantic fields covered by the subdatabase; however, there are considerable
differences in the proportion of words from this source. Whilst both the mean and
the median of the number of loans in the different semantic fields are just above
20%, indicating that there is no significant imbalance caused by individual semantic
categories, the percentage of loanwords from these languages ranges from a mere
3.5% (Sense perception) to 58% (Modern world). For the latter category, it has al-
ready been mentioned above that most of the meanings it contains did not exist
prior to their introduction via the cultures corresponding to the respective donor
languages. A similar explanation can be given for the category Animals with 22% of
(primarily French) educated loanwords, which contains a considerable number of
animals that are not indigenous to the Romanian-speaking territory and for which
the terms were therefore borrowed when Romanian speakers first became aware of
their existence via French; examples are c"mil" camel, elefant elephant or cangur
kangaroo.
With a total of 27% of loanwords from these source languages, the semantic
field Clothing and grooming is also strongly affected. In this category, Romanian has
2
Some loanwords from German could also be included in this class, but it is not always clear whether
a word was borrowed from German as a language representing a cultural model or through to con-
tact with the German minority in Romania.
borrowed from virtually all contact languages over the centuries, incorporating cor-
responding words for new types of clothing items introduced by or through the
respective population or culture. A considerable number of words from French and
Italian have been added as (near) synonyms or hyponyms to existing ones, reflecting
an orientation towards emulating western European fashion, including the use of
the corresponding vocabulary. For instance, beret" and basc" have been borrowed
from French to denote specific types of cap, complementing the existing #apc"
(borrowed from Bulgarian %apka) and c"ciul" (cognate with Albanian ksul), both of
which also refer to caps. Similarly, gheat" boot was borrowed from Italian ghetta,
joining the default word for boot, cism" (borrowed from Hungarian csizma) and
ciubot" (borrowed from Ukrainian &oboty).
Similarly, in the areas Cognition and Emotions and values, with 26% and 29% of
words borrowed from languages serving as cultural models, many of the loans are
synonyms of existing words, incorporated into Romanian due to the fashion and
prestige associated with the use of educated loans. Terms from these languages
associated with the areas Law and Hunting and warfare (35% and 27% respectively),
on the other hand, were borrowed together with the new objects and concepts they
denoted, due to fundamental changes to the legal system in post-Ottoman Romania
and significant advances in military technology and strategy, respectively.
At the other end of the scale, the semantic fields Sense perception (3%), Quantity
(6%), The house (7%), Food and drink (7%), and Kinship (8%) have incorporated far
fewer loans from languages representing cultural models during the past two centu-
ries. Whilst Sense perception and Quantity appear to be generally more resistant to
borrowing pressure, with only 13.9% and 17.7% of words in the respective catego-
ries identifiable as loans from any external source, the number of items borrowed
from other source languages in the fields Kinship and Food and drink indicates that
these semantic categories were specifically more resistant to borrowing from west-
ern European languages serving as cultural models.
The second most significant source of loanwords in Romanian are the Slavic lan-
guages, due to prolonged, close contact between Romanians and the Slavic peoples
that moved into the area. Within the subdatabase, 8.4% of all words are borrowed
from (South) Slavic with no particular regional provenance, 5.4% from Bulgarian
and/or Serbian, 0.7% from Ukrainian, and a few isolated items from Russian and
Polish. The total percentage of loanwords from Slavic sources is 14.6%, approxi-
mately one seventh of the words in the database.
It has already been mentioned above that a considerable proportion of Slavic
loans entered Romanian through the use of Old Church Slavonic as the language of
religion, particularly in the semantic field Religion and belief (25%), but also in the
category Social and political relations. It is not always easy to decide whether a loan-
word was borrowed from Slavic primarily via the religious domain or due to the
general prolonged and close contact between the two languages due to cohabitation
during several centuries. In the category Emotions and values, for instance, 17% are
borrowed from South Slavic including Bulgarian; it is likely that the borrowing of
words such as mil" pity from Slavic mil$ is due to its use in everyday conversation
as well as specifically religious contexts.
Other semantic fields that show a significant number of loans from Slavic are
Speech and language (17%), Basic actions and technology (14%), Time (12%), The
physical world (12%), Possession (12%), Motion (11%), The house (10%), and
Warfare and hunting (10%).
Loanwords from Slavic have, in many cases, replaced inherited words even where
their meanings have been continually present since Roman times; for instance, nisip
sand is borrowed from Bulgarian nasip, z"pad" and om"t (both snow) from Slavic
zapad$ and omet$, all but replacing the respective inherited synonyms arin" and
nea, which are nowadays restricted to regional and poetic use. In a similar way, izvor
spring from Slavic izvor$ has ousted the inherited fntn" in this meaning; under
the influence of Italian, French and German, fntn" has nowadays shifted its
meaning to fountain.
In other cases, loans from Slavic have filled lexical gaps; the most visible example
is da yes, a notion that could not be rendered by any single word in Latin. In nu-
merous instances, however, Slavic loanwords co-exist with synonymous inherited
words; in many cases, there is little or no discernible difference in meaning or in
usage frequency. Borrowing of an exact synonym from Slavic can eventually lead to
semantic differentiation; inherited timp and borrowed vreme (from Slavic vr'men),
synonyms referring to both time and weather, for instance, show an incipient
semantic split, with vreme increasingly becoming the more common choice for
weather.
On the other hand, a number of Slavic loanwords have fallen victim to a strong
th
re-latinisation process since the 19 century. Thus, the Slavic loanword cern black
has disappeared from modern usage, ousted by the inherited synonym negru.
th
A different development can be observed with the pair german (a 19 century
learned loan from Latin) and neam! from Slavic n'mici (both German), which are
used virtually synonymously in everyday conversation, though some speakers feel it
is inappropriate to call a German neam! to his face, even though it is not generally
perceived as a disrespectful term. This example shows that synonym pairs created
by borrowing can come to contain complex and unpredictable semantic and socio-
pragmatic nuances that go beyond their lexical meaning.
In antonym pairs with one element borrowed from Slavic, there is an intriguing
tendency for the Slavic word to be the one with more positive connotations. Exam-
ples are to love vs. to hate (a iubi from Slavic ljubiti vs. inherited a ur), friend vs.
enemy (prieten borrowed from Slavic prijatel vs. du#man borrowed from Turkish
d#man), and yes vs. no (da borrowed from Slavic vs. inherited nu).
About 2% of the words in the subdatabase for Romanian are borrowed from Turk-
ish. Whilst these loanwords account for a far smaller proportion of Romanian
vocabulary than those discussed in the previous sections, it is still a considerable
amount of the lexical stock. It is perhaps noteworthy that not a single Romanian
verb in the database is borrowed from Turkish. By far the largest impact can be
observed in the lexical field The house, with 10% of the vocabulary in this category
borrowed from Turkish. The majority of these loanwords can be attributed to in-
novations and improvements in construction, furniture, tools etc., as well as
fashions, introduced from the Ottoman Empire. Some examples are chirpici adobe
from Turkish kerpi, chio#c garden house, kiosk from Turkish k#k, geam window
from Turkish cam, sob" stove from Turkish soba, hogeag chimney from Turkish
ocak. Many of these words were widely borrowed throughout the Balkan areas of
the Ottoman Empire and have cognate loanforms in neighboring languages; Turk-
ish cam, for example, has also been borrowed into Greek as %&'( (dzmi), into
Albanian xham [)am] and Bulgarian d(am.
Other semantic fields with a relative large proportion of loans from Turkish are
Agriculture and vegetation (4.7%), Food and drink (4.5%), and Clothing and grooming
(4.5%). In a way similar to the previous examples of words borrowed from Turkish,
these loans are typically linked to the introduction of the corresponding objects as a
result of Turkish influence through the close political and commercial links with
the Ottoman Empire.
In some cases, Turkish loanwords are synonymous with inherited items, with no
discernible semantic difference; one of the more visible examples due to its central
nature as part of Romanian rural culture is cioban shepherd from Turkish oban, a
synonym of the inherited word pastor.
There are also some intriguing antonym pairs in which the Turkish loanword
tends to be the element with more negative connotations. Examples are clean vs.
dirty (inherited curat vs. murdar borrowed from Turkish) and friend vs. enemy
(prieten borrowed from Slavic prijatel vs. du#man borrowed from Turkish d#man).
1.7% of the words in the database are borrowed from Greek, with the highest pro-
portion in the semantic fields Clothing and grooming (6.1%), Modern world (5.8%),
The house (5.7%), and Religion and belief (4.5%). Many of the loans from Greek
entered the language in a way similar to the Turkish loans, due to contact and trade
between Greeks and Romanians within the Ottoman area of influence; religious
terminology can be attributed to the shared Orthodox Christian background.
A similar percentage, 1.6%, is borrowed from Hungarian. The largest number
of loans, in the category Social and political relations (6.5%) can be attributed to the
fact that Transylvania, a large section of the Romanian-speaking territory, was un-
th th
der Hungarian influence or rule between the 11 and the 20 century, and as a
result, social relations were influenced by Hungarian concepts. Even relatively fun-
damental social concepts such as gazd" host from Hungarian gazda and a se ntlni
to meet from Hungarian talni were incorporated into the common Romanian
vocabulary and are not limited to the territories that were under Hungarian influ-
ence. Other lexical fields with a relatively high proportion of Hungarian loanwords
are Clothing and grooming (4.5%), Speech and language (4.5%), and The house
(4.3%). In all these semantic categories, new words were introduced together with
the corresponding culturally specific objects and concepts from Hungarian.
Hungarian loanwords also fill genuine lexical gaps; in Romanian, there is normally
no distinction made between leg and foot, both of them rendered by picior. To
make specific reference to foot without the leg, the word lab", borrowed from
Hungarian lb.
Loanwords from Albanian or closely related pre-Latin substrate languages ac-
count for only 1% of the vocabulary in the database. Only a limited number of
semantic fields contain loans from this source: The physical world (4.8%), Kinship
(3.2%), Agriculture and vegetation (2.8%), Animals (2.7%), Social and political
relations (1.6%), The house (1.4%), The body (1.2%), and Clothing and grooming
(1.1%). Virtually all loanwords from Albanian/substrate fall into the areas of family,
farming, and basic living. The fact that these words survived the process of lan-
guage shift to Latin indicates that terms and concepts belonging to these areas of
life were so deeply rooted in the culture that they continued to be used during and
even after the shift to Latin.
5. Integration of loanwords
5.1. Phonological integration
Generally, the majority of loanwords are not subject to a great deal of phonological
change, largely due to the relative large phonological inventory and tolerant phono-
tactics of Romanian. It is likely that this phonological tolerance is itself, at least in
part, due to the continual influx of borrowed words from various source languages.
A number of phonological features of Romanian may have emerged as a result of
contact and large-scale borrowing; the central vowel /!/, for instance, may have
entered Romanian from Slavic (Hall 1974: 73), though this claim is disputed by
Petrucci (1999: 6069).
Certain voiced word-initial consonant clusters such as /zdr-/ are also likely to
have developed due to borrowing from Slavic, as they regularly occur in loanwords
from Slavic beginning with /s*dr-/.
Romanian Slavic
a zdrobi to annihilate s$drobiti
zdrav"n strong, healthy s$dravin$
zdrean!" rag s$dran$
The phonological development leading to the creation of this cluster is not limited
to Romanian; cognate forms such as Serbian zdravo healthy show that this is a
more widespread, regional process. However, the occurrence of the same initial
consonant cluster in Romanian words with unknown origin may be an extension of
this phonological sequence beyond the originally borrowed items, indicating that
the cluster has become a fully integrated part of the language.
Some vowel distinctions in the source languages that do not exist in Romanian
have led to a change in vowel quality (and quantity where applicable) in the bor-
rowed Romanian word. For instance, in ora# town from Hungarian vros [va+,o-],
3
the quantity distinction is not preserved , and in du#man enemy from Turkish
d#man, the Turkish close front rounded vowel /y/ is replaced by a close back
rounded vowel /u/. Similarly, the close-mid front rounded vowel [] has been re-
placed by the close-mid back rounded vowel [o] in chio#c garden house from
Turkish k#k, but a trace of the original vowel is preserved in the fact that the pre-
ceding /k/ is followed by a palatal glide, which is a regular development affecting
the sequence /k/ + close-mid front vowel. On the other hand, the more recently
borrowed foen [fn] hairdryer (from German Fhn) preserves the original vowel
almost unchanged, thereby effectively adding a new vowel phoneme to the
Romanian inventory.
Some phonological patterns are perceived to be typical of loanwords from a particu-
lar donor language. A example of this are nouns ending.in -ea when indefinite and
in -eaua when definite, which are perceived by most Romanians to be of Turkish
origin. In many cases, this is accurate.
However, this pattern is, in fact, indigenous, as shown in the inherited word for
star.
3
Whether the metathesis of the two vowels is linked to the original difference in quantity is unclear.
Almost all loanwords are morphologically fully integrated; most verbs borrowed
from Slavic are incorporated into the conjugation of infinitives ending in /-i/, due
to the similarity to the original ending in /-iti/. Verbs from other Romance lan-
guages and learned Latin borrowings are generally integrated into the conjugations
corresponding to those the respective nouns belong to in their source languages.
However, a set of borrowed verbs ending in /-r/ in Romanian, most of them of
Slavic origin, have developed a modified paradigm that can be analyzed as a newly
created conjugation (Schulte 2005).
Nouns and adjectives are also generally fully integrated, following one of the
various declension patterns available. The original gender of nouns is generally re-
tained, and inanimate objects that do not have a gender in their source language
usually receive neuter gender and morphology, which implies masculine agreement
in the singular and feminine agreement in the plural. Examples are gref grapefruit
from English grapefruit and pix ballpoint pen, probably borrowed via English from
the brand name Bic, with the respective plural forms grefuri/grefe and pixuri requir-
ing feminine agreement. The reanalysis of borrowed plurals ending in /-s/ as
singular forms and the subsequent addition of the plural morpheme /-uri/ is com-
monly found, especially with loans from English, as in cips (SG) from English
(potato) chips, with the plural cipsuri.
As will be briefly discussed in 6, Romanian has borrowed a large number of
South Slavic affixes, many of which appear on loanwords from South Slavic but
have become productive affixes in Romanian. As a result, the presence of originally
Slavic derivational morphology on loanwords does not set them apart from other
Romanian words.
6. Grammatical borrowing
As a member of the well-known Balkan convergence area, Romanian has developed
a considerable number of morphological and syntactic structures in parallel with the
other member languages of the Balkan Sprachbund. For some of the features shared
among these languages, it is impossible to determine in which direction they were
borrowed, or whether they are simply the result of joint development. These in-
clude postposition/suffixation of the definite article, a case system in which the
genitive and the dative have merged, an analytic comparative, the formation of the
numerals eleven to nineteen with a preposition meaning on or over, following the
pattern one-over-ten, two-over-ten etc., formation of an analytic future tense with
a verb originally meaning to want, and the use of an empty imperative verb
(Greek ide, Turkish haydi, Bulgarian/Romanian haide), used to encourage the
addressee to go ahead with an unspecified but pragmatically obvious action; in Bul-
garian and Romanian, haide can receive morphological person inflection (Romanian
haidem (1PL) and haide!i (2PL)).
For other features shared by the Balkan languages, a specific source language can be
pinpointed. For instance the widespread tendency to use finite subordinate clauses
instead of infinitival clauses even in cases of subject coreference is generally attrib-
uted to the merger of the infinitive and subjunctive in Greek, subsequently
spreading across the Balkan Peninsula.
Other features of Romanian morpho-syntax can be attributed directly to specific
source languages. The use of an obligatory attribute agreement marker is a feature
of Albanian, and presumably of pre-Latin substrate languages in Romania, that has
been borrowed into Romanian, albeit with greater restrictions on its usage.
Among the morphological features clearly borrowed from South Slavic are the
use of a vocative in /-o/ for feminine nouns ending in /-a/ in their nominative
form, the loss of the final syllable of the infinitive, and a large set of derivational
morphemes directly borrowed from South Slavic (du Nay 1996: 102108).
Finally, in more recent times, an increased use of the infinitive since the second
th
half of the 19 century may, at least in part, be attributable to the strong influence
of French (Close 1974: 227). However, this is not a complete innovation; the
French model is likely to have acted as a reinforcement of an incipient resurgence of
the infinitive in certain constructions, especially in prepositional adverbial clauses
(Schulte 2007: 308316).
7. Conclusion
Having borrowed from a considerable number of languages over the centuries, Ro-
manian can serve as an example of a language with a high degree of lexical
permeability. Borrowing has taken place in a number of very distinct types of con-
tact situations, ranging from cohabitation and population mix on the one end of
the scale (e.g. South Slavic) to predominantly indirect contact (e.g. French). De-
spite the fundamentally different nature of these contact situations, both have
provided a large amount of morphologically and phonologically fully integrated
lexical material.
Whilst the differences between contact situation types affect the number of
loanwords from particular languages in certain semantic categories, borrowed items
are found in all areas of the lexicon. Loanwords are not only used for objects and
concepts for which there was no indigenous word, but have also been introduced as
synonyms for existing words; in some cases this has led to the creation of multiple
synonyms from a number of different donor languages. Borrowed synonyms often
coexist with little or no discernible meaning difference, though their availability is
sometimes exploited to make subtle semantic or pragmatic distinctions. Some
loanwords are typically associated with a particular register, but generally even
words from high-status source languages are used in colloquial speech, which shows
that they have been fully integrated into the language.
The continuous addition of lexical material from various source languages over
the centuries means that a large proportion of the vocabulary of present-day Roma-
nian is not inherited from Latin; in some semantic areas, loanwords far outnumber
inherited ones. Even relatively basic words denoting continually present meanings,
such as features of the natural environment, are frequently borrowed. Whilst it
might therefore be argued that Romanian is a language with a hybrid vocabulary,
the large number of words borrowed from other Romance languages over the last
two centuries nevertheless gives its lexicon a distinctly Romance appearance.
References
Cihac, Alexandru. 1879. Dictionnaire dtymologie daco-romane. Vol. 2: lmemts slaves,
magyars, turcs, grecs-moderne et albanais. Frankfurt (Main): St. Goar.
Cior"nescu, Alexandru [Alejandro]. 1966. Diccionario etimolgico rumano. La Laguna:
Biblioteca Filolgica & Madrid: Gredos.
Close, Elizabeth. 1974. The development of modern Rumanian. Oxford: Oxford University
Press.
Conev, Benju. 1921. Ezikovni vzaimnosti me(du b"lgari i rum"ni. Sofia: Sofia University
Press.
Corominas, Joan. 1961. Breve diccionario etimolgico de la lengua castellana. Madrid: Gredos.
Coteanu, Ion et al. 1998. Dic!ionarul explicativ al limbii romne. 2nd edn. Bucharest: Univers
Enciclopedic/Academia Romn" Institutul de Linvistic" Iorgu Iordan.
Diculescu, Constantin. 192426. Elementele vechi grece!ti din limba romn". Dacoromania
IV:394516.
du Nay, Andr. 1996. The Origins of the Rumanians: The early history of the Rumanian
language. Toronto/Buffalo: Matthias Corvinus Publishing.
Fiedler, Wilfried & Klosi, Ardian. 1997. Wrterbuch Deutsch-Albanisch. Berlin/Mnchen:
Langenscheidt.
Grebe, Paul et al. 1963. Duden Etymologie: Herkunftswrterbuch der deutschen Sprache [Duden
Etymology: Etymological dictionary of the German language]. Mannheim:
Bibliographisches Institut (Dudenverlag).
Gruji$, Branislav. 1998. D(epni re&nik nema&ko-srpski. Belgrade: Janus.
Hall Jr. Robert A. 1974. External History of the Romance Languages. New York: Elsevier.
Ha!deu, Bogdan Petriceu. 1877. Columna lui Traian VIII (1877).
Ha!deu, Bogdan Petriceu. 1879. Cuvinte den b"trni. Vol. 1. Bucharest.
Ha!deu, Bogdan Petriceu. 1883. Columna lui Traian XIV (1883).
Isb"!escu, Mihai. 1995. Dic!ionar german-romn: 60.000 de cuvinte. Bucharest: Teora.
Levi#chi, Leon & Banta!, Andrei. 1992. Dic!ionar englez-romn. 70.000 de cuvinte.
Bucharest: Teora.
McClure, Erica F. 1976. Ethnoanatomy in a multilingual community: An analysis of
semantic change. American Ethnologist 3(3):525542.
Loanword Appendix
Thraco-Dacian or Albanian om"t snow brici razor
vreme weather, time colib" hut
m"gur" mountain, hill
jar embers ograd" yard, court
mal shore
j"ratic embers stlp doorpost, post,
pru river, stream
fl"c"u young man pole
mlac" swamp
nevast" wife z"vor latch, door-
abur fog, steam bolt
maic" mother
scrum ash grind" beam
bab" grandmother,
copil child brn" beam
old woman
mo# grandfather, tab"r" camp
odrasle descendants
old man
izlaz pasture hrle! spade
moa#" grandmother,
coco# cock, rooster lopat" shovel
old woman
scoic" shell a s"di to sow
!ap he-goat
c"mil" camel a cosi to mow
mnz foal, colt
prepeli!" quail coas" sickle, scythe
c"pu#" tick
veveri!" squirrel ov"z oats
#oprl" lizard
bivol buffalo orez rice
cioc beak
trup body a munci work
ceaf" nape of neck
obraz face, cheek a ncovoia to bend
groap" grave
a clipi to blink a lovi to strike, hit,
c"ciul" hat, cap
beat, to kick
vatr" fireplace glezn" ankle
a (d)obor to cut down
gard fence pizd" vagina, vulva
topor axe, ax
copac tree a se trezi to wake up
tesl" adze
brad fir a omor to kill
a tr to pull
c"tun village strv carcass
a tescui to squeeze
ran" wound, sore
(South) Slavic a cl"di to build
leac medicine
a zidi to build
praf dust otrav" poison
clei glue
pr"pastie cliff, precipice gol naked, empty
covaci blacksmith
ostrov island a n"bu#i to choke
nicoval" anvil
val wave a pr"ji to roast, fry
cositor tin, tinplate
izvor spring, well cle#te tongs
sticl" glass, bottle
smrc swamp m"slin" olive
plas" netbag
mla#tin" swamp ulei oil
dalt" chisel
stnc" stone, rock psl" felt
a (r")suci to turn, to
vijelie storm mantie cloak
twist
v"zduh air podoab" ornament,
a t"v"li to roll
pcl" fog adornment
a cl"ti(na) to shake
z"pad" snow perie brush
a stropi to splash