SELIM 
Journal of the Spanish Society for Mediaeval English Language and Literature

Revista de la Sociedad Española de Lengua y Literatura Inglesa Medieval

Nº 
Oviedo, 

Edited by - Dirigida por 
Trinidad Guzmán & S. G. Fernández-Corugedo

Universidad de Oviedo &
Sociedad Española de Lengua y Literatura Inglesa Medieval

Oviedo – Murcia 



Isabel de la Cruz Cabanillas, Selim  (): –ISSN: 1132–631X

SEMANTIC PRIMES IN OLD ENGLISH: 
A PRELIMINARY STUDY OF DESCRIPTORS¹

Abstract
The aim of this paper is to apply the methodology of semantic primes by Goddard and 
Wierzbicka () to Old English in order to check whether it represents a suitable 
theoretical and methodological amework for the lexical and semantic study of this period. 
This constitutes a preliminary analysis of the semantic primes grouped as Descriptors: 
BIG/SMALL. The group is discussed taking into account a sample of texts provided by 
The Helsinki Corpus of English Texts and supplemented by the information contained in 
The Dictionary of Old English Corpus. The main sources of information on Old English 
definitions are A Thesaurus of Old English by Roberts and Kay () and A Concise Anglo-
Saxon Dictionary by Clark Hall (). The article attempts at being just a first approach 
to the topic, which could be further developed and extended to other semantic categories. 
Keywords: Old English, semantic primes, linguistic corpus.

Resumen
El objetivo del presente artículo es aplicar la teoría de los primitivos semánticos de 
Goddard y Wierzbicka al inglés antiguo para comprobar si representa un marco teórico 
y metodológico adecuado para el estudio léxico-semántico del periodo. Éste constituye 
un análisis preliminar de los primitivos semánticos agrupados bajo la denominación de 
Descriptors: BIG/SMALL. El grupo se analiza teniendo en cuenta una muestra de textos 
extraídos del Helsinki Corpus of English Texts. El artículo intenta ser un primer acercamiento 
al tema, que podría seguir investigándose en el futuro, a medida que se amplíe el corpus de 
análisis a través de la información proporcionada por el Dictionary of Old English Corpus. 
Las principales fuentes de información para las definiciones en inglés antiguo proceden de 
The Oxford English Dictionary, A Thesaurus of Old English de Roberts y Kay, así como A 
Concise Anglo-Saxon Dictionary de Clark Hall. Palabras clave: inglés antiguo, primitivos 
semánticos, corpus lingüísticos.

I 

A lthough two opposing views on linguistic facts have been alternating throughout the history of linguistics and even if some aspects are culture-dependent, as proposed by 
¹ The research reported here is part of the project “Base de datos dinámica 
online de la morfología derivativa del inglés antiguo” (“Dynamic online database 
of Old English derivative morphology”), sponsored by the Spanish Ministry of 
Education, Reference HUM2005-07651-C02-02/FILO.



38

Isabel de la Cruz Cabanillas

Selim  ()

relativism, some other linguistic facts seem to obey to universally 
valid rules.² In fact, a uniform language was already reported in 
the Bible referring to the Tower of Babel. Therefore, the quest to 
find out what is invariable and shared by all languages has a long 
history and om different theoretical positions both philosophers 
and linguists have strived along centuries to discover these universal 
features common to all human languages. The debate was centred 
not only on the finding of universals but also on their own existence, 
inasmuch as some scholars deny the actual existence of universal 
properties common to all languages. However, proponents of 
linguistic primitives are found among several schools of thought: 
realism in the early Middle Ages, nominalism and conceptualism in 
the Renaissance period, rationalism in the th century as well as 
other movements and individuals in the next two centuries who tried 
to discover the relationship between the abstract universal entities 
and the particular languages that embody them, the best method 
to approach linguistic primitives and how to define them. Some of 
the ideas defended within the philosophical ameworks were put 
into practice by linguists om the th century onwards, although 
the objectives were not always alike. Thus, the Neogrammarians 
developed a whole network of connections between languages based 
on their genealogy or common origin; typological linguists focused 
on grouping languages according to the morphological constituents 
basically, while generativists centred their research on the syntactic 
structure of the different systems. Even if the approaches and 
the goals were diverse, the investigation on linguistic universals 
benefited om the advance in linguistics, as efforts were made to 
acquire a deeper knowledge of a huge amount of languages and to 

² For a revision of the philosophical and linguistic origin and the development 
of language universals, as well as the different standpoints within this amework, 
see Mairal (), Mairal & Gil () and Moure (), among others.



39

Semantic primes in Old English

Selim  ()

compare them to find out about their origin, their morphemes, 
their syntactic patterns or their divergences.

In the th century, especially in its second half, there was a 
renewed impetus in the search for universal linguistic properties. 
The debate on those general linguistic common traits was the focal 
point of investigation in two crucial conferences held at Dobbs 
Ferry (), “Conference on Universals of Language”, organised 
by Greenberg and at Austin (), “Symposium on Universals 
in Linguistic Theory”, by Bach and Harms. As a result of the 
arguments exposed in these two events, studies to test hypothetical 
sets of universals across a number of genetically and typologically 
diverse languages were undertaken. They were applied to the 
different planes of language analysis: phonology, morphology, syntax 
and lexis and semantics. Within the latter field, seminal work has 
been done on the part of Goddard and Wierzbicka. Thus, following 
the tenets of universal grammar, Wierzbicka (: ) defends 
the idea that semantic primes or fundamental human concepts are 
universal and innate. Likewise, the Natural Semantic Metalanguage 
amework assumes that every natural language can be used as a 
metalanguage, which makes it unnecessary to resort to abstract 
semantic predicates to account for the meanings of a given language. 
Although there have been several previous attempts in which the 
set of hypothetical primitives was considerably expanded om one 
version to the next one, in its latest version the Natural Semantic 
Metalanguage includes (Goddard : ) the following:

Substantives•  (, , , , /, )
Determiners (• ,  , )
Quantifiers (• , , , , /)
Evaluators (• , )
Descriptors (• , )
Mental predicates (• , , , , , )



40

Isabel de la Cruz Cabanillas

Selim  ()

Speech (• , , )
Actions, events and movement (• , , )
Existence and possession (•  , )
Live and death (• , )
Time (• /, , , ,   ,   
,   )
Space (• /, , , , , , , 
)
Logical concepts (• ,  , , , )
Intensifier, augmentor (• , )
Taxonomy, partonomy (•  ,  )
Similarity (• )

Following the tenets rendered in Goddard and Wierzbicka () 
and Wierzbicka (), this article aims at contributing to the model 
of universal grammar by carrying out the application of the Natural 
Semantic Metalanguage Research Program (henceforth ) to 
Old English. Martín Arista & Martín de la Rosa () already 
applied the model to the analysis of the semantic primes of Old 
English that belong to the classes of substantives, determiners and 
quantifiers. The common characteristic to the three categories is that, 
being grammatical words, their combinations are far more restricted 
than those of lexical classes. As the latter article constitutes the first 
attempt to put the  into practice in Old English, the present 
research follows similar techniques and methods, although on this 
occasion I will concentrate on the analysis of descriptors /, 
a lexical class. In choosing this specific approach my purpose is 
twofold: Firstly, to check whether it represents a suitable theoretical 
and methodological amework for the lexical and semantic study 
of the period. Secondly, to establish the ‘hyperonym⒮’ among all 
possible equivalents of  and  extracted om the Old English 
Thesaurus by Roberts and Kay ().



41

Semantic primes in Old English

Selim  ()

M  
Before discussing the exponents of the semantic primitives it is 
necessary to ponder on two preliminary issues:

First of all, do all languages have units to express the concepts 
of  and ? The evidence presented by different scholars, 
such as Goddard and Wierzbicka,³ seems to suggest that they do. 
As one cannot resort to native speakers to confirm this claim in 
Old English, lexicographic works are one of the most valuable 
information sources. As Wierzbicka (: ) suggests the right 
way to proceed is by identiing “in the language the local exponents 
of the universal concepts in question, with all their allomorphs and 
allolexes (i.e., lexical variants) and with the relevant grammatical 
ames”. When one looks up for the equivalents of  in Roberts 
and Kay (: -), it reads as follows:⁴

.... Greatness, bigness, size: micelnes, 
micelu

. Greatness, bigness: greatnes, gryto

. Fatness, bulkiness: fætnes

. Something huge, a very great mass: ormæte

. Large, big, broad: brad, micel, wid

.. Large, capacious, ample: sid

. (Of a ship) ample, with a large hold: widfæþme

³ Some of their more relevant contributions to the establishment of the theory 
and the progressive development of the different sets of universals can be 
consulted in the reference section. See, for instance, Goddard (; a & 
b), Goddard & Wierzbicka (; ) and Wierzbicka ( & ). In 
some of these publications, the reader will also find the application of the model 
to other languages.

⁴ Although long vowels are marked with a macron on the paper version of the 
Thesaurus, the online version offers the possibility of searching with or without 
length-mark. As the Helsinki Corpus and the The Dictionary of Old English Corpus 
show no vowel length, this feature is omitted all through the article.



42

Isabel de la Cruz Cabanillas

Selim  ()

. Massive, bulky, huge: great
… (Of things) plump, full-bodied, subtantial: fætt
. Considerable: gehycglice, ungehwæde
. Not little, great: unlytel
. Very great, exceedingly great: swiþlic
. Huge, immense, enormous: eacencræig, ormæte, 

ormætlic, ungefog, ungefoglic, unmæte, 
unmætlice, ungemet, uþmæte

. Giant, gigantic: entisc, eotenisc

. Of varying size: mismicel

The first thing that strikes the reader by having a look at the 
elements is the fact that the word big is not included, as it is not 
documented until the end of th century in writers of Northumbria 
and north Lincolnshire. The Oxford English Dictionary relates its 
location to a possible Norse origin. Likewise, large was not part of 
the word-stock yet, as it is first attested in the Middle English period. 
Apart om that, the second striking aspect is the great amount 
of lexical units that can be used. According to Wierzbicka, when 
trying to identi a proper evaluator (: ) “difficulties involved 
in identiing these concepts cross-linguistically are due more to 
the superabundance of plausible candidates than to their absence”. 
Although she is referring to evaluators /, the great number 
of synonyms or quasi-synonyms documented in Old English for the 
descriptors / seems to pose the same problem. It follows 
om here that the variation in use requires further investigation, 
as can be seen om the glance at the exponents of Old English 
semantic primitive of  (Roberts and Kay, : -):

... Littleness, smallness: gehwædnes, 
lytelnes



43

Semantic primes in Old English

Selim  ()

. Little, a small amount: hwæthwugu, hwega, hwon, 
hwugudæl, lyt, lytel, lythwon, medmicel, 
tohwega.

. A particle, small piece, jot: corn, egl, grot, lyttuc, 
mot, prica, pricele, spot, wloh

. (Of amount) small, little: hwæde, lytel, medeme, 
smæl, unginne, unmicel

.. Little, small, slight: hwonic, ieþelic, mæte, medemlic, 
medmicel

. Very little: forlytel, lytel

.. Least: læst

. Somewhat, a little: æthwara, æthwega, be sumum dæle, 
hwæthwara, hwæthwugu, hwæthwugununges, 
hwene, hwon, sumdæl, sume dæle, sumes, sum 
on dæle, tohwega

. Slightly, little: hwæthwugu, hwon, hwonlice, leohtlice, 
lyt, lytl, lythwon, lytle, lytlum, medemlice

The abundance of terms for both concepts ( and ) makes 
it difficult to work out which one is the hyperonym among all the 
possible candidates. Two selecting criteria can be applied here: on 
the one hand, the equency of occurrences; on the other, the range 
of different collocations a word shows. In order to apply the first 
criterion the group of descriptors is discussed taking into account 
a sample of texts included in The Helsinki Corpus of English Texts. 
The use of written corpora is not ee om criticism, as has been 
pointed out by different scholars (Görlach, : ). If written 
corpora present a series of deficiencies, historical ones show more 
specific constraints. Some of limitations that are usually adduced for 
the study of historical texts and that are particularly related to the 
topic of representativeness of the corpus are:



44

Isabel de la Cruz Cabanillas

Selim  ()

They just include written texts with their limitations, that is, () 
written texts are not representative of all registers, genres, age, 
sex, or social condition of speakers.
Modern readers do not have access to every text produced at a () 
specific period. They are restricted to some types of texts that 
contemporary readers considered it was worth copying.
The survival of the original texts is oen arbitrary and by chance, () 
which implies the data are not complete as a consequence of the 
random preservation. This means that there is a random selection 
of texts.

All these constraints may minimize the representativeness of the 
sample. The validity of the data has also been questioned on the 
part of some scholars. The written sources compiled in corpora 
are limited in size. It follows om here that the sample may be 
considered invalid because of its provenance and reduced size 
(Schneider, : -).

As the Helsinki Corpus has been widely contrasted and is 
worldwide accepted, it is taken for granted all the requirements for 
a suitable selection of the different sources have been met. Thus, 
the equency parameter is based on the data retrieved by using 
this corpus. Nonetheless, the results will show that there are some 
elements that are not present in the Helsinki Corpus. That is why 
The Dictionary of Old English Corpus (henceforth DOEC) has also 
been consulted to try to solve some of the problems encountered 
when using the former corpus. The results of both will be contrasted 
to see if there are any significant differences. 

The second criterion is the variety of collocation a particular 
item shows. Faber & Mairal (: ), calling this feature the 
Principle of Lexical Iconicity, refer to the fact that “the greater the 
semantic coverage of a lexeme is, the greater its syntactic variation”. 
Or put it the other way round, Cortes & Mairal (: ) define it 



45

Semantic primes in Old English

Selim  ()

as “the greater the syntactic coverage of a lexical unit, the higher its 
position in the semantic hierarchy within a given subdomain”. One 
will be able to find out about the collocations of the items by consulting 
Roberts and Kay (), available online now, which enables users to 
search for phrases in Old English. In addition, those collocations 
that appear thrice or more in the Helsinki Corpus are also selected. 
The information contained in this source will be completed by that 
provided by A Concise Anglo-Saxon Dictionary (Clark Hall, ).

D   
The combination of both criteria, equency and range of collocations, 
should give us reliable clues to be able to determine which elements 
might have been used as superordinates, although it may turn 
out difficult to decide which of several was the real hyperonym.  
Nonetheless, before having a look at some of the examples, it is 
necessary to comment on the researcher’s need to go through the 
data personally. Obviously the automatic search enables the scholar 
to handle a wide variety of data, but also poses some problems:

First of all, as we are dealing with an untagged corpus, no 
distinction is made between homonymy among the different word 
classes. Thus, sid and wid can be both an adjective or an adverb. In 
fact, very oen the adverbial combination side and wide or vice versa 
wide and side is read through the texts. Side can even be a noun, as 
well. In addition, by browsing an adjective like mæte, we find plenty 
of occurrences which correspond to a verb in third person singular; 
most of them show the construction Gyf mon/man mæte… (‘If one 
considers…’). Even more, if the search is done with the adjective in 
nominative we can have no occurrences, as happens with medemlic. 
That is why the option with final asterisk is preferred to find all the 
possible inflections and that is how medemlice is found. However, 
when the option mæt* is used, non desired examples like mæterne are 



46

Isabel de la Cruz Cabanillas

Selim  ()

retrieved as well. It follows om here that the data must be revised 
in order to filter the ones scholars really need for their analysis.

The researcher must also be aware of language variation. 
Although the variety of written forms is not so wide as in the 
Middle English period, we find some alternants, like in the case of 
micel that is found  times either as quantifier or descriptor in the 
nominative case. Furthermore, if the search is carried out by mycel 
with ⟨y⟩ instead of ⟨i⟩  tokens are displayed.

Finally the presence of some items is not so pervasive in the 
Helsinki Corpus as shown in Roberts and Kay (). In fact, even if 
the latter work provides several phrases for great, just  hints will 
be retrieved om the Helsinki Corpus, out of which  correspond to 
the name of the author, Pope Gregory the Great. This fact must be 
taken into account, as the search will include the title of the work 
as well as the name of the author.

From the whole set of items provided by Roberts and Kay () 
those that could be interpreted as quantifiers have been disregarded, 
because Martín Arista & Martín de la Rosa handled those data 
in their article. Nevertheless, elements such as hwon or hwæthwugu 
or micel are found not only as quantifiers, but also as descriptors. 
Therefore, all the items that could function as descriptors have 
been analysed concentrating on those cases where they function 
as adjectives and not as other word categories. Likewise, nouns 
implying size have also been disregarded, as  and  will 
basically appear as adjectives in attributive function.

Thus, for the Old English exponents of   entries provided 
by Roberts and Kay () have been revised. Out of these, no 
occurrences were found for mismicel, swiþlic, ormætlic, unmætlice, 
uþmæte, ungefoglic, gehycglice and ungehwæde. The rest of the 
elements show the following equency and collocations either in 



47

Semantic primes in Old English

Selim  ()

Roberts and Kay () or in the Helsinki Corpus (henceforth HC), 
as displayed in Table :

Table . OE exponents of BIG, their frequency and collocations

BI
G

Item Number of occurrences Phrases
 – > 

entisc !

micel ! micel lic (‘elephantiasis’)
micel wundor (‘great wonder’, HC)
micel folc (‘big crowd’, HC)mismicel !

great ! great sealt (‘coarse salt’)
greate wyrt (‘meadow saffron’)
seo greate banwyrt (‘unidentified plant’)
þa greatan netlan (‘nettle’)

brad ! brad hand (‘big hand’)
brad sweord (‘broad sword’)
brad þistel (‘big thistle’)

eotenisc !

swiþlic !

ormæte !

ormætlic !

unmæte !

unmætlice !

uþmæte !

wid ! (on/to) widan feore (‘eternity, for ever’)
sid ! side rice (‘large kingdom’, HC)
unlytel !

ungemet !

eacencræftig !

ungefog !

ungefoglic !

gehycglice !

ungehwæde !



48

Isabel de la Cruz Cabanillas

Selim  ()

Similarly, for the Old English exponents of   items have 
been revised, out of which no information was retrieved on any of 
the following: forlytel, hwonic, ieþelic, lytl, as can be seen in Table 
:

Table . OE exponents of SMALL, their frequency and collocations

SM
A

L
L

Item Number of occurrences Phrases
 – > 

ieþelic !

mæte !

medmicel !

medmicle !

medemlic(e) !

forlytel !

lyt !

lytel ! lytel forca (‘little fork’)
lytle ! se lytla finger (‘the little finger’)
lytlum ! lytlum fæce (‘a small interval of time’, HC)
lythwon !

læst !  se læsta finger (‘the little finger’)
hwæthwugu !

hwon !

hwonic !

hwonlice !

leohtlice !

The lack of results about the mentioned units can only be taken as 
an indication of the low equency of the terms, although obviously 
not being included in the compilation does not mean that the words 
were non-existent. A quick glance at the rest will make readers 
aware of the fact that there are several items that are etymologically 



49

Semantic primes in Old English

Selim  ()

related in each set. Thus, regarding the primitive embodied by  
the bigger group is made up of the stem mæt with different affixes: 
ormæte, ormætlic, unmæte, unmætlice, uþmæte. In the  group 
we find lytel and its variants (lyt, lytle, lytlum, lythwon), medmicel 
(medmicle, medemlice), hwon (hwonlic, hwonlice) and then mæte, 
læst, hwæthwugu and leohtlice. From these the lytel family clearly 
outnumbers the occurrences found for the other items.

If a comparison between both descriptors is established, it can 
be observed that  words out of  show no occurrences for ;  
are just represented by very few instances ranging om  to  and 
finally just  exponents of Old English  are found more than  
times in the corpus. In the case of , no hints are retrieved for 
 of the  items;  lexical units show a equency ranging between 
 and  times in the Helsinki Corpus and the pending elements,  
in total, are found on more than  occasions. There is not always 
a coincidence between the most widely recorded words in the 
corpus used and those whose phases are mentioned as common 
in Old English by Roberts & Kay (), but both criteria can be 
combined to strive to figure out which ones could have been used as 
hyperonyms and are, because of that, more equent. 

These results highlight the insufficiency of the Helsinki Corpus 
to provide occurrences for all the searched items on this occasion. 
Therefore, it seems necessary to look for complementary data in a 
bigger corpus. As the Dictionary of Old English Corpus comprises 
at least one copy of each text surviving in Old English, it will 
surely meet all the requirements to be a suitable complement to 
the study. In fact, the retrieval of data om the DOEC proves 
to be an important supplement. The number of tokens extracted 
outnumbers those found in the HC in most cases. As a matter of 
fact, most of the items show a higher equency in the DOEC than 



50

Isabel de la Cruz Cabanillas

Selim  ()

in the HC, but the real difference is made in relation to the absence 
of some items.

Regarding the primitive  nearly all the items absent om 
the HC are now found in the DOEC (mismicel, swiþlic, ormætlic, 
unmætlice, uþmæte, ungefoglic, and ungehwæde) with the exception 
of gehycglic. For instance, Roberts and Kay () mention that this 
word appeared just once in Anglo-Saxon documents. Furthermore, 
the term is not registered by Clark Hall () as such, although 
hycglic—without the prefix—appears in Die Dialoge Gregors den 
Grossen  : . The only explanation for its absence om the 
DOEC would be that the variant text of Gregory’s Dialogues 
included in the DOEC introduced a synonym of hycglic, rather than 
this very word.

On the contrary, all the three elements absent for  in the 
HC are now present in the DOEC: namely forlytel, hwonlic, ieþelic. 
However, some of the problems have not been solved by the use 
of the DOEC: for instance, being also an untagged corpus, no 
distinction is established between parts of speech. Thus, cases of 
homonymy between word categories must still be worked out by the 
researcher. This task is even harder as the number of occurrences 
is much higher. Not only is this tendency observed in cases like the 
ones mentioned above for lytel or micel and other related forms, but 
quite a simple form like brad in the HC, where  tokens are found, 
appears with  hints in the DOEC. Thus, the DOEC helps in 
providing more reliable data regarding some items, but presents 
similar limitations to those found in the HC as well. 

As the outcomes obtained through the use of the DOEC do 
not alter significantly those retrieved om the HC, the presence 
of the semantic primes will now be illustrated by using examples 
om those that have  or more instances as descriptors in the 
Helsinki Corpus. The exponent that displays the greatest number 



51

Semantic primes in Old English

Selim  ()

of occurrences is micel. Even if micel and lytle can be found as 
quantifiers, as can be seen in the following passage om the Anglo-
Saxon Chronicle, most of the occurrences happen to be adjectives 
in attributive position. Some of them would even be controversial, 
for instance micle aþas sworon, where micle could be interpreted as 
a quantifier and translated as ‘swore many oaths’ or as a descriptor 
and be rendered as ‘swore great oaths’. The latter is the option 
chosen by both Savage (: ) and Swanton (: -) for 
the following sentence:

& micle aþas sworon() 
Some other instances of micel are clearly quantifiers as in: 

& geridon Wesseaxna lond & gesæton micel þæs () 
folces
‘and they rode over the Wessex land and sat up 
many of the people’

From the same extract examples with lytle as quantifier () and 
descriptor () are taken:

& he lytle werede unieþelice æfter wudum for() 
‘and he defended a little with greater difficulty 
through the wood’

þæs on Eastron worhte ælfred cyning lytle werede () 
geweorc
‘at Easter king Aled built a little fort’

The other exponents for  showing more than  occurrences 
are brad, ormæte, unmæte, sid, wid, unlytel and ungemet. Let’s see an 
example of each one:

ofer brad brimu Brytene soht() 
‘Brytene sought over the big waves’



52

Isabel de la Cruz Cabanillas

Selim  ()

gif ormæte hungor cymð() 
‘If big hunger comes’

com werod unmæte() 
‘a big multitude came’

we widefeorh weorcum hlodun geond sidne grund() 
‘we built (with) works through the ample ground 
for a long time’.

Example () could illustrate both sid and wid; the latter oen appears 
with feorh ‘life, time span’ inasmuch as it is sometimes considered an 
adverb translated as ‘always’.

to miclan bryne wæter unlytel() 
‘to a big fire, big/much water’

þæt is asolcennyss, ðæt is modes swærniss and () 
ungemetegod slapulniss
‘that is laziness, that is sluggishness of spirit and 
excessive somnolence’.

By going through the various sentences retrieved for , it has been 
observed that some of the descriptors function as intensifiers for 
other adjectives implying size, in the same way extremely, enormously 
or immensely can be used in present-day English. In fact, Clark Hall 
() records ormæte and ungemet both as adjectives and adverbs. 
With this function are found in the Helsinki Corpus:

se mona is ormæte brad() 
‘the moon is immensely large’

winter bringeð weder ungemet cald() 
‘the winter brings extremely cold weather’.

Regarding the specific set of items for , apart om lytel/lytle 
and other etymologically related units like lytlum and lyt, the only 



53

Semantic primes in Old English

Selim  ()

ones that are found  times or more are læst—that is also related to 
lytle as being its superlative—, hwæthwugu and hwon.

Babylonia, seo ðe mæst wæs & ærest ealra burga, seo () 
is nu læst
‘Babylon, which was the largest and the first of 
all towns, is now the smallest’.

Hwæthwugu is documented as noun, adjective, pronoun and adverb 
in Clark Hall (). Thus, many of the occurrences will go under 
any of the other word classes. Besides, it is one of the few that can 
present graphic variants, since the last part of the compound can be 
recorded as -hwigu, -hugu, -hwega, -hwegu or -hwygu. An example of 
hwæthugu as descriptor is found in ():

& gedyde hwæthugu getæse() 
‘and did little profit’.

Finally, one of the items showing the greater number of occurrences 
is hwon, as it can be a form of the interrogative hwa, found in adverbial 
phrases like to hwon, for hwon meaning ‘why’; this is probably the 
most widely found case. Besides, it can be an adjective, a noun and 
an adverb. Most of the hints supplied are either as an inflected form 
of hwa or as an adverb. As a descriptor is found in:

geond þas eorðan æghwær sindon hiora gelican hwon () 
ymbspræce
‘through the earth everywhere is their same little 
speech’.

C  
In the introduction section, the goals of the present research 
were set up: Firstly, to check whether the Natural Semantic 
Metalanguage Research Program represented a suitable theoretical 



54

Isabel de la Cruz Cabanillas

Selim  ()

and methodological amework for the lexical and semantic study of 
Old English and secondly, to establish the ‘hyperonym⒮’ among 
all possible equivalents of  and  extracted om the Old 
English Thesaurus by Roberts and Kay () by means of the 
interpretation of the data retrieved om the Helsinki Corpus of 
English Texts and the Dictionary of Old English Corpus. 

Regarding the first objective of the investigation, it can be 
concluded that, although the Natural Semantic Metalanguage  
Research Program has dealt with various natural languages, English 
is probably one to which little attention has been paid. Thus, one of 
the initial aims of the article was to test the validity of the Natural 
Semantic Metalanguage Research Program as a useful tool in a 
former stage of English, namely Old English. Even if this piece of 
work attempted at being just a first approach to the topic, there is 
no doubt that the findings and proposals of the Natural Semantic 
Metalanguage Research Program may be tested in Old English. 
With this brief analysis I hope to have demonstrated that semantic 
primes represent a useful tool for the lexical and semantic analysis 
of Old English inasmuch as it can certainly be applied to other levels 
of linguistic analysis, such as morphology and syntax, for instance. 
Furthermore, it seems convenient to extend the research to other 
categories established by Goddard and Wierzbicka () in order 
to find the exponents in Old English and to present the specific 
peculiarities this stage of the language shows.

Although the first objective was clearly achieved, there are some 
issues related to the second objective that need further research, 
such as which of the terms can clearly be considered hyperonyms. 
As a matter of fact, this second objective was tested and developed 
in the discussion of the data retrieved om Roberts and Kay () 
and the two corpus used: Helsinki Corpus and The Dictionary of Old 
English Corpus. Even if some methodological problems were solved 



55

Semantic primes in Old English

Selim  ()

by handling the data provided by the latter, such as the retrieval of 
some items absent in the former, there are still some deficiencies 
that prevail. Probably the most salient one is the fact that, being 
both untagged corpora, there is a need on the part of the researcher 
to filter the data in order to select the right information for the 
investigation.

Isabel de la Cruz Cabanillas
Universidad de Alcalá de Henares

R
Primary sources
Clark Hall, J. R. : A Concise Anglo-Saxon Dictionary. Cambridge 

University Press, Cambridge.

The Dictionary of Old English Corpus. Online version available at: 
http://ets.umdl.umich.edu/o/oec/ 

Oxford English Diccionary. Online version available at: http://
dictionary.oed.com

Rissanen, M. & O. Ihalainen. : The Helsinki Corpus of English 
Texts. Diachrnic and Dialectal. Universidad de Helsinki, 
Helsinki.

Roberts, J. & C. Kay. : A Thesaurus of Old English. King’s 
College London Medieval Studies, London and online version 
available at: http://libra.englang.arts.gla.ac.uk/oethesaurus

Secondary sources
The Anglo-Saxon Chronicles. : Translated and collated by A. 

Savage. Crescent Books, New York.



56

Isabel de la Cruz Cabanillas

Selim  ()

The Anglo-Saxon Chronicles. : Translated and edited by M. 
Swanton. Phoenix Press, London.

Cortés, F. & R. Mairal : A preliminary design for a syntactic 
dictionary of Old English on semantic principles. Díaz Vera, J. 
E. ed. A Changing World of Words. Studies in English Historical 
Lexicography, Lexicology and Semantics. Rodopi, Amsterdam: 
-.

Faber, P. & R. Mairal : Constructing a lexicon of English verbs. 
Mouton de Gruyter Berlin.

Goddard, C. : Semantic theory and semantic universals. 
Goddard, C. & A. Wierzbicka eds. Semantic and Lexical 
Universals. Theory and Empirical Findings. John Benjamins, 
Amsterdam: -.

Goddard, C. a: The Search for the Shared Semantic Core of 
All Languages. Goddard, C. & A. Wierzbicka eds. Meaning 
and Universal Grammar. Theory and Empirical Findings 
(Volume I). John Benjamins, Amsterdam: -.

Goddard, C. b: The On-going Development of the NSM 
Research Program. Goddard, C. & A. Wierzbicka eds. 
Meaning and Universal Grammar. Theory and Empirical 
Findings (Volume II). John Benjamins, Amsterdam: -.

Goddard, C & A. Wierzbicka : Introducing lexical primitives. 
Goddard, C. & A. Wierzbicka eds. Semantic and Lexical 
Universals. Theory and Empirical Findings. John Benjamins, 
Amsterdam: -.

Goddard, C. & A. Wierzbicka. : Semantic Primes and Universal 
Grammar. Goddard, C. & A. Wierzbicka eds. Meaning and 
Universal Grammar. Theory and Empirical Findings (Volume 
I). John Benjamins, Amsterdam: -.



57

Semantic primes in Old English

Selim  ()

Görlach, M. : Corpus problems of text collections: linguistic 
aspects of the canon. Studies in the History of the English 
Language. Carl Winter, Heidelberg: -.

Martín Arista, F. J. & Mª V. Martín de la Rosa : Old English 
Semantic Primes: Substantives, Determiners and Quantifiers. 
ATLANTIS /: -.

Mairal Usón, R. : Reflexiones en torno a la noción de ‘universal 
lingüístico’. Fuertes Oliver, P. coord. Lengua y sociedad: 
Investigaciones recientes en Lingüística Aplicada. Universidad de 
Valladolid, Valladolid: -.

Mairal, R. & J. Gil eds. : En torno a los universales lingüísticos. 
Cambridge University Press, Cambridge.

Moure, T. : Universales del lenguaje y linguo-diversidad. Ariel, 
Barcelona.

Schneider, E. : Investigationg Variation and Change in Written 
Documents. Chambers, J. K., P. J. Trudgil and N. Schilling-
Estes eds. The Handbook of Language Variation and Change. 
Blackwell, Oxford: -.

Wierzbicka, A. : Semantic primitives: The expanding set. 
Quaderni di Semantica /: -.

Wierzbicka, A. : Semantic Primes Across Languages: A 
Critical Review. Goddard, C. & A. Wierzbicka eds. Semantic 
and Lexical Universals. Theory and Empirical Findings. John 
Benjamins, Amsterdam: -.

Wierzbicka, A. : Semantics: Primes and Universals. Oxford 
University Press, Oxford.

Wierzbicka, A. : Anchoring linguistic typology in universal 
semantic primes. Linguistic typology /: -.



58

Isabel de la Cruz Cabanillas

Selim  ()

Wierzbicka, A. : Semantic Primes and Linguistic Typology. 
Goddard, C. & A. Wierzbicka eds. Meaning and Universal 
Grammar. Theory and Empirical Findings (Volume II). John 
Benjamins, Amsterdam: -.

!

Received  Feb ; revision received  May ; accepted  Jun 