E. C.

QUIGGIN MEMORIAL LECTURES 14

Odd Einar Haugen

‘So that the writing may be less and quicker, and the
parchment last longer’: The orthographic reform of the
Old Icelandic First Grammatical Treatise

DEPARTMENT OF ANGLO-SAXON, NORSE AND CELTIC

UNIVERSITY OF CAMBRIDGE
Edmund Crosby Quiggin (1875-1920) was the first teacher of Celtic in the
University of Cambridge, as well as being a Germanist. His extraordinarily
comprehensive vision of Celtic studies offered an integrated approach to the
subject: his combination of philological, literary, and historical approaches
paralleled those which his older contemporary, H. M. Chadwick, had
already demonstrated in his studies of Anglo-Saxon England and which the
Department of Anglo-Saxon, Norse and Celtic continues to seek to emulate.
The Department has wished to commemorate Dr Quiggin’s contribution by
establishing in his name, and with the support of his family, an annual
lecture and a series of pamphlets. The focus initially was on the sources for
Mediaeval Gaelic History. Since 2006 the Quiggin Memorial Lecture is on
any aspect of Celtic and/or Germanic textual culture taught in the
Department.

‘So that the writing may be less and quicker, and the parchment last longer’:
The orthographic reform of the Old Icelandic First Grammatical Treatise

© Odd Einar Haugen

First published 2012 by the Department of Anglo-Saxon, Norse and Celtic,

University of Cambridge, 9 West Road, Cambridge, CB3 9DP.

ISBN 978-0-9571862-4-8
ISSN 1353-5722

Set in Andron by Odd Einar Haugen

Printed by the Reprographics Centre, University of Cambridge.

 E. C. QUIGGIN MEMORIAL LECTURES 14

Odd Einar Haugen

‘So that the writing may be less and quicker, and the
parchment last longer’: The orthographic reform of the
Old Icelandic First Grammatical Treatise

DEPARTMENT OF ANGLO-SAXON, NORSE AND CELTIC

UNIVERSITY OF CAMBRIDGE
‘So that the writing may be less and quicker, and the parchment last longer’
The orthographic reform of the Icelandic First Grammatical Treatise

ODD EINAR HAUGEN

University of Bergen

The First Grammatical Treatise (FGT) is one of four grammatical treatises

preserved in the Icelandic manuscript AM 242 fol, along with several other texts,
notably the Prose Edda by Snorri Sturluson. This manuscript is often referred to
as Codex Wormianus (or, in Icelandic, Ormsbók) after its owner in the seven-
teenth century, the Danish physician and antiquary Ole Worm (1588–1655). It is
now in the Arnamagnæan Collection in Copenhagen.
The first two of the grammatical treatises deal with grammar in the modern
sense, especially orthography, while the third and the fourth discuss rhetorical
matters. In the medieval concept of trivium, grammar, rhetoric, and logic all
belonged to ars grammatica, but of the four grammatical treatises in Codex
Wormianus, only the first two would be referred to as grammatical today. The
first treatise, which is the one to be discussed here, is deservedly the best known
of the four. It is not a long text, comprising seven pages in the manuscript, or
about 4,000 words. However, it is a very compact text and it is obviously the
work of a gifted grammarian. It has often been claimed that this is the first text
using the concept of minimal pairs in establishing the most suitable orthography
of a language. This means that it predates the modern concept of minimal pairs
introduced by the Prague phonologists in the 1930s. However, being written in a
little known language on the periphery of Europe, it had no impact on the
European grammatical tradition. It remained a unique contribution, only to be
rediscovered and appreciated anew in modern times.
The FGT in Codex Wormianus is the only copy preserved of this work. The
original was by all accounts written in the middle of the twelfth century. 1
Between the twelfth century original and the fourteenth century copy there are
an unknown number of intermediate copies. Some readings point to there being
at least one intermediate copy, but this is in all likelihood only a small part of the

1
Einar Haugen, First Grammatical Treatise, 2nd ed., London: Longman, 1972, seems to
favour the traditional dating to 1130–1140 (p. 78), while Hreinn Benediktsson, The First
Grammatical Treatise, Reykjavík: Institute of Nordic Linguistics, 1972, more cautiously
suggests 1125–1175 (pp. 31–33).
2 E.C. Quiggin Memorial Lectures

Fig. 1. Codex Wormianus with the opening of the FGT. Note that the scribe has left a space for
the initial ‘J’. It was quite common that initials were drawn by someone other than the scribe,
and in a surprising number of cases, the initial was never made. Here, the scribe has indicated
the initial by adding a small ‘J’ in the square set aside for it. Copenhagen, The Arnamagnæan
Collection, AM 242 fol, p. 84 (= fol. 42v).
‘So that the writing may be less and quicker’ 3

manuscript tradition which once existed.2 According to common estimates, as

many as 90 per cent of medieval manuscripts have been lost.3

Fig. 2. Two manuscripts – one scribe. To the left is an extract from AM 227 fol (Stjórn), fol.
38r, and to the right an extract from AM 242 fol (Codex Wormianus), p. 25 (= fol. 13r). There
can hardly be any doubt that these lines are the product of the same scribe, who is also known
to have written several other manuscripts. Both manuscripts have been dated to around 1350.
In this image, the size of the script in the two manuscripts has been scaled for easier com-
parison.

In other words, we know the text of the FGT in its later guise, but we do not
know what it originally looked like. Even if we assume that the text was copied
faithfully, some new readings must have been introduced along the way. We
know that the Icelandic language changed on a number of counts over the two
hundred years from the mid-twelfth to the mid-fourteenth century. Conse-

2
See Hreinn Benediktsson, The First Grammatical Treatise, pp. 22–23.
3
See e.g. John L. Cisne, ‘How Science Survived: Medieval Manuscripts’ “Demography”
and Classic Texts’ Extinction,’ Science 307 (2005): 1305–1307, with further references.
4 E.C. Quiggin Memorial Lectures

quently, the orthography changed, too. We also know that the style of writing
altered markedly in the same period.
In this article, I will try to answer two well-nigh impossible questions about
the original state of the FGT. I do realise that the answers to these questions are
indeed uncertain. However, there is a fascination in trying to unveil the secrets of
a lost manuscript, even if the quest only covers the brief period from the mid-
twelfth to the mid-fourteenth centuries. Thus my questions are as follows: (1) In
what kind of script was the original committed to parchment? (2) If the FGT had
been written in the orthography it recommends, what would the orthography
have looked like?

The script of the original

The first question is the easiest one, which makes it a convenient starting point.
There is no doubt that the appearance of the text in Codex Wormianus (W) is
far from what the original must have looked like. The script of W is a fully
developed Gothic script as witnessed in many Icelandic manuscripts of the time.
Some of them were in fact written by the same scribe, as can be seen from Fig. 2,
where the writing of AM 242 fol is compared with the Bible translation Stjórn in
AM 227 fol.
The scribe of W wrote the manuscript in the style of his day, not trying to
capture any of the older traits. To the best of my knowledge, copying the style of
an exemplar was first introduced in the Old Norse manuscript tradition in the
sixteenth century. A prime example is the copy made by bishop Jens Nilssøn in
1567 of the medieval manuscript Jǫfraskinna, now lost except for a few leaves.
Jens Nilssøn’s transcription is displayed in Fig. 3. Anyone familiar with medieval
Nordic script will recognize the medieval traits in this writing, such as the
Insular shape of ‹ f › and, above all, the large number of abbreviations. This is an
example of someone trying to copy a text – not only with respect to its contents,
but also with respect to its palaeographical form.
My first question is primarily a palaeographical question, and in trying to
answer it, one has to look at the style of contemporary Icelandic manuscripts, i.e.
manuscripts that were produced in the middle of the twelfth century.4 There is a

4
A good overview is Hreinn Benediktsson, Early Icelandic Script, Reykjavík: Manuscript
Institute of Iceland, 1965. This contains facsimiles and transcriptions of all important Ice-
landic manuscripts up to the end of the thirteenth century. A very useful supplement is
‘So that the writing may be less and quicker’ 5

Fig. 3. The kings’ saga Jǫfraskinna (1300–1325) in a transcription of the now mostly lost medi-
eval exemplar. Jens Nilssøn (1538–1600) made this copy in 1567, and he obviously tried to
copy his source as exactly as possible, including all abbreviation marks. Copenhagen, The
Arnamagnæan Collection, AM 37 fol., fol. 150r, l. 8–14.

striking difference between the earliest Norwegian and Icelandic manuscripts in

the fact that the earliest Norwegian manuscripts are written in a Carolingian
minuscule with distinct Insular traits, while the earliest Icelandic manuscripts
were written in a pure Carolingian script. Around 1200, the Carolingian script
gave way to the style which is now commonly termed Proto-Gothic script. Sub-
sequently, a fully developed Gothic script came into use.5 Thus, assuming that
the FGT was conceived and committed to parchment for the first time in the
mid-twelfth century, we can be fairly sure that it was written in a pure Caro-
lingian style.
The preserved Icelandic manuscript material from the twelfth century is not
large. In the handbooks by Hreinn Benediktsson and Guðvarður Már Gunn-
laugsson (cf. note 4) there are five specimens which may be regarded as Caro-
lingian, i.e. the single leaf of Reykjaholtsmáldagi (the first 14 lines dated to c.
1130–1150), AM 237 a fol, two leaves of homilies (c. 1150), AM 315 d fol, two

Guðvarður Már Gunnlaugsson, Sýnisbók, 2nd ed., Reykjavík: Stofnun Árna Magnússonar,
2007. The Sýnisbók contains a wide range of facsimiles in full colour (supplemented by
transcriptions) of Icelandic specimens from around 1100 to around 1900.
5
On the term proto-gothic, see Albert Derolez, The Palaeography of Gothic Manuscript Books,
Cambridge: Cambridge University Press, 2003, ch. 3.
6 E.C. Quiggin Memorial Lectures

Fig. 4. Reykjaholtsmáldagi containing a list of the possessions of the Reykjaholt church. A

single leaf of 36 lines written over a period of time from the first half of the twelfth century to
c. 1300. Reykjavík, Þjóðskjalasafni Íslands. The first 14 lines (displayed here) have been dated
to 1130–1150 (Guðvarður Már Gunnlaugsson, Sýnisbók, 2nd ed., 2007, p. 12).

leaves of the law Grágás (c. 1150–1175), AM 674 a 4°, Elucidarius, 33 leaves (1150–
1200) and AM 673 b 4°, Plácítúsdrápa, 5 leaves (c. 1200).6 After having recon-
sidered the earliest Icelandic material, Guðvarður Már Gunnlaugsson believes
that only three of these are truly Carolingian, viz. Reykjaholtsmáldagi (the first 14
lines), AM 237 a fol and AM 674 a 4°.7

6
These manuscripts are all kept in Reykjavík, The Arnamagnæan Collection.
7
As for Norwegian manuscripts, he is of the opinion that only one Norwegian specimen is
to be regarded as fully Carolingian, viz. GKS 1347 4°, fol. 62v. This is a single page listing
properties belonging to the Munkeliv monastery in Bergen, which has been written on the
last leaf of a Gospel manuscript in Latin. His views are developed in the article, ‘Caro-
lingian and Proto-Gothic Script in Norway and Iceland,’ in the forthcoming Studies in
Memory of Lilli Gjerløw, edited by Espen Karlsen.
‘So that the writing may be less and quicker’ 7

Reykjaholtsmáldagi is commonly regarded as the oldest preserved Icelandic

manuscript leaf. As can be seen from the facsimile in Fig. 4, it is an open script
with a fair amount of space between the individual letters and between the lines.
It differs from early Norwegian script in that it has ‹ þ › in all positions of the
word (instead of ‹ þ › and ‹ ð ›), and, in addition, it does not contain the Insular
letter forms of ‹ f ›, ‹ v › and ‹ r ›. Note that in the known Icelandic specimens from
the twelfth century there is considerable variation in the format of the books, the
width of the column and the size of the letters. While Elucidarius in AM 674 a 4°
is a rather small codex with comparatively large letters, Reykjaholtsmáldagi has
smaller letters and approximately double the column width of Elucidarius.
The first 14 lines of Reykjaholtsmáldagi, recently dated to c. 1130–1150, is
contemporaneous with or possibly somewhat earlier than the FGT.8 As men-
tioned in note 1 above, the FGT has been dated to 1125–1175 or, more narrowly,
to 1130–1140. The script in these lines is thus the prime candidate for a palaeo-
graphical reconstruction of the original FGT. After I had tried to reconstruct the
orthography (the procedure for this will be discussed below), the calligrapher Bas
Vlam was given the task of actually reproducing the script.9 Bas Vlam has done
similar work over the last few years; the first was a reconstruction of the lost half
of a Norwegian chirograph, a letter from Aga, Ullensvang in Hardanger, 26 May
1293.10 The next was a reproduction of the script in the Old Norwegian homily
book (c. 1200 or somewhat later).11 The present reproduction is his first venture
into Icelandic script.
As can be seen from Fig. 5, the reconstruction is very close to the exemplar.
For this to be possible, the calligrapher has to estimate the angle between the
letters and the base line, and the width of the pen nib. Furthermore, he has to
deconstruct the script of the exemplar into its constituent strokes; it is not only a
question of trying to copy the finished letter, but of building it with the right
number of strokes in the right sequence.12 When looking at the two specimens in
Fig. 5, one has to conclude that they look as if they have been written by the
same scribe – although the latter one is, as one would expect, the more regular.

8
See Guðvarður Már Gunnlaugsson, Sýnisbók, p. 12.
9
His work can be accessed at http://www.kalligraf.no/.
10
Bergen, University Library, Diplomsamlingen. The reconstruction is reproduced in colour
in the chapter ‘Diplomer, lover og jordebøker’ by Jon Gunnar Jørgensen, in Handbok i
norrøn filologi, ed. Odd Einar Haugen, Bergen: Fagbokforlaget, 2012.
11
Copenhagen, The Arnamagnæan Collection, AM 619 4°. See the article by Bas Vlam and
his reproduction in Vår eldste bok, eds. Odd Einar Haugen and Åslaug Ommundsen, Oslo:
Novus, 2010, pp. 101–113.
12
This procedure is explained and illustrated in his article in Vår eldste bok.
8 E.C. Quiggin Memorial Lectures

Fig. 5. Two scribes, almost 900 years apart. The first six lines of Reykjaholtsmáldagi (c. 1130–
1150), and six lines of the reconstructed script of the First Grammatical Treatise by Bas Vlam
(2011), based on the script of the Reykjaholtsmáldagi.

Deciding on the style of the script is only the first step in the reconstruction
of the text; it is by no means an unexpected conclusion that the FGT was most
probably written in a Carolingian style. The script must also be applied to a
specific orthography; in other words: form requires content.

What did the orthography look like?

In this article, I am working from the hypothesis that the FGT was originally
written in its own recommended orthography. In the concluding remarks of this
article, I will briefly discuss whether this is a reasonable hypothesis, but let us for
the sake of argument accept it as a point of departure. It is only to be expected
that this will lead to a text looking quite different from the one in W. Of W, one
can safely refer to the old adage that you should do as I say, not as I do.
The first grammarian takes as his starting point the Latin alphabet as he
knew it. He does not suggest changing any of its letters, but he argues in favour
of a somewhat different selection of letters, and suggests additions among the
vowels as well as among the consonants. With regards to the vowels, Ancient
‘So that the writing may be less and quicker’ 9

Nordic had the five vowels /i/, /e/, /a/, /o/ and /u/, as had Latin. As a conse-
quence of syncopation and i- and u-mutation, four new vowels were added to
this inventory, as shown in Fig. 6. Representing the vowels appropriately is in
many ways the first grammarian’s primary concern.

Fig. 6. The vowel system of Old Icelandic in the twelfth century. The four new vowels appear
in circles. They are the result of i- and u-mutation.

The main points in the recommended orthography of the FGT are the following:

1. There should be one and only one letter for each sound in the language.

This is perhaps the most important principle of the FGT, although it is not for-
mulated as explicitly as here. The fact that the orthography of the FGT is sup-
posed to be in a 1 : 1 correspondence with the underlying phonological system
has led to its being used as the Archimedean point in the analysis of Icelandic
(and, in general, Old Norse) phonology.13 Most linguists seem to believe that it
offers a ‘perfect fit’, i.e. that a text written in the recommended orthography of
the FGT would be a near-orthophonic text. I add ‘near’ since there are a few
exceptions to this rule, namely that he allows ‹ x › for the sequence /ks/, which is
redundant. Also, his use of a special character for /ng/, namely the strange-
looking ‹  › of Codex Wormianus, is questionable, and raises doubts about an
unequivocal phoneme–grapheme correspondence. Finally, it has been pointed
out that short nasal vowels were not really in opposition to short oral vowels,
since the nasality of short vowels, e.g. in words like land and menn, were

13
A recent example is the chapter on Old Icelandic and Old Norwegian language, ‘Gammal-
islandsk og gammalnorsk språk’, by Jan Ragnar Hagland in Handbok i norrøn filologi,
Bergen, 2012.
10 E.C. Quiggin Memorial Lectures

derivable from the nasal context. Even so, among the orthophonic orthographies
of the world, past and present, the one recommended by the first grammarian
surely ranks high on the list.

2. There are short and long vowels in the language, and they should be disting-
uished by an acute accent (stryk).

The analysis of vowel quantity is correct, and the recommendation of using

accents is appropriate, although not wholly original. Diacritical marks were spo-
radically used to indicate length in Latin manuscripts, either an apex or an acute,
but it is not clear to what extent diacritical marks were used in the writings
known to the first grammarian. Although usage was not consistent, accents were
employed to indicate long vowels in some thirteenth-century Icelandic manu-
scripts, e.g. AM 645 4° (part 1, ca. 1220; part 2, c. 1225–1250). Accents were used
sporadically in Old Norwegian manuscripts, too, such as the Old Norwegian
homily book, AM 619 4° (c. 1200 or shortly after). It is not clear whether this
usage can be attributed to the FGT. In other respects, it looks as if Icelandic
script was influenced from Norway, not the other way round. For example, the
early Icelandic practice of using ‹ þ › in all positions (correct from a phonemic
point of view) was replaced be the Norwegian practice of using ‹ þ › initially and
‹ ð › in other positions (a subphonemic variation).

3. There are oral and nasal vowels in the language, and the nasal vowels should
be marked by a superscript dot.

This is perhaps the most remarkable analysis and recommendation in the trea-
tise. There are no traces of this practice in Icelandic or Norwegian orthography.
The superscript dot was used in somewhat younger Icelandic manuscripts, but
mostly over consonants to indicate length. It seems likely that Norwegian and,
hence, Icelandic had nasal vowels for several centuries after the period of syn-
copation 500–700 AD. In this period, short vowels were syncopated unless they
were nasal; thus Ancient Nordic *landa n. ‘land’ was syncopated to Old Norse
land, while in an infinitive like *takan ‘take’, the final n was dropped, but it left
the preceding short a with a nasal timbre, and this vowel was in fact not syn-
copated, cf. Old Norse taka. In other words, there must have been a distinction
between oral and nasal vowels at the time, which is also borne out by runic
orthography. After the loss of initial /j/ in the syncopation period, e.g. /ja:ra/ >
/a:r/, Old Norse ár n. ‘year’, the old j rune, ᛃ , with an added stave, ᛡ, was
‘So that the writing may be less and quicker’ 11

shifted to represent the oral /a/ (it was later changed into ᛆ ), while the old oral a
rune, ᚫ , was taken into use for the nasal /ã/, somewhat later in other forms, e.g.
ᚮ .14 We cannot be sure about the nasality of other vowels, and there are no traces
of a further distinction in runic script. However, it should be added that the runic
orthography in the Viking Age was highly restricted. There were only four
runes, i ᛁ , u ᚢ , ą ᚮ and a ᛆ, and they represented a total of nine vowels, which in
addition were distinguished by length and nasality. In other words, the lack of
runes for nasal vowels – apart from ą ᚮ vs. a ᛆ – cannot count as an argument
against the existence of nasality in the other vowels. From the Norwegian runic
inscriptions it seems that Norwegian lost nasality in the eleventh century, while
the FGT leads us to believe that Icelandic kept nasality somewhat longer. If we
assume that nasality was lost in Icelandic in the twelfth century, it is no longer
surprising that there is no trace of it in later orthography. It seems that the FGT
caught a linguistic phenomenon just as it was disappearing.

Fig. 7. The vowel system of the FGT. Each vowel can be short and oral (no diacritics), long
and oral (acute accent), short and nasal (superscript dot) or long and nasal (acute accent and
superscript dot).

Fig. 7 shows the Icelandic vowel system at the time of the FGT with its nine
vowels. Following the distinction between oral and nasal, and short and long,

14
One should have expected the opposite to happen, but the reason seems to be that the
initial sound in the traditional rune name were regarded as defining for the sound repre-
sented by the rune (the acrophonic principle). For this reason, the old a rune in */ansuʀ/,
Old Norse áss n. ‘heathen god’, was the best candidate for the nasal /ã/ since it was in fact
nasalised in the rune name, */an-/. In */ja:ra/ there was no nasality, and when the initial
/j/ was dropped, cf. Old Norse ár n., the j rune was no longer a good candidate for /j/
(which in fact became represented by the rune for /i/), but it came to be used for the
initial sound of the new rune name, */a-/, i.e. the oral /a/.
12 E.C. Quiggin Memorial Lectures

vowels, the FGT concluded with no less than (9 × 2 × 2) = 36 distinctions. In

this figure, the vowels are displayed with the two diacritical marks suggested by
the first grammarian, the acute accent for length and the dot for nasality.

4. There are short and long consonants in the language, and the long consonants
should be rendered by small capitals.

This is another remarkable recommendation in the FGT and it is a wholly origi-

nal one. A similar practice is to the best of my knowledge not known from other
vernacular orthographies in Europe. This practice was also taken up to some
extent in later Icelandic orthography, especially for the consonants ‹ ɴ ›, ‹ ʀ › and
‹ ɢ ›. Small capitals as such were not the invention of the FGT, but in other ortho-
graphies, e.g. in Norwegian writing, they were only used ornamentally.15

Fig. 8. The consonant system of the FGT. Long consonants are represented by small capitals.
Three consonants, ‹ þ ›, ‹ h › and ‹ ŋ › were always short and have no corresponding capital form.
Note that /k/ is represented by ‹ c › and /s/ by the tall form ‹ ſ ›. The table is simplied with
respect to /g/, since the short variant represented by ‹ g › could be both plosive and fricative,
while the long variant represented by ‹ ɢ › could only be plosive. The table is also simplified
with respect to /h/; it may be regarded as an unvoiced vowel, but with respect to distribution
it should be analysed as a consonant, especially in clusters like ‹ hl- ›, ‹ hn- › and ‹ hr- › (the /h/
was dropped in Norwegian, but retained in Icelandic).

15
An example of this can be found in a Norwegian charter from 1225, cf. the facsimile in
Didrik Arup Seip, Palæografi, B: Norge og Island, Oslo 1954, p. 31, Fig. 7, l. 5, where we
find the oblique form of Maria spelt ‹ Maʀiu ›. Here, /r/ can not possibly be long.
‘So that the writing may be less and quicker’ 13

The standard option was to render long consonants by geminates, i.e. by

doubling, such as ‹ aptann ›, ‹ herra › and ‹ liggja ›. This is in fact the modern Scan-
dinavian way of doing it. It could be argued that capitals were an uneconomical
way of dealing with consonantal length. Why did the grammarian not recom-
mend a single diacritical mark for length, the acute accent? In other words, why
spell /si:n/ as ‹ sín ›, but /sin:/ as ‹ siɴ › rather than as ‹ siń ›? In modern parlance,
we might counter that this would leave too little redundancy in the orthography,
and that a degree of redundancy is needed in any channel of communication.
Also, the first grammarian might add that either solution would be equal with
regards to syntagmatic length; it is as space-saving to write ‹ siɴ › as it is to write
‹ siń ›. Thus, his major concern to save parchment was not invalidated.
Fig. 8 displays the consonant system of the FGT. Here, short consonants are
displayed with their usual Latin form, while the long consonants are displayed
with small capitals. This is the one recommendation where the first grammarian
really lives up to his dictum of making the script become quicker and shorter.

5. The semivowel /j/ should be rendered with ‹ i › (but ‹ e › in rising diphtongs),

and /v/ with ‹ u ›.

The FGT does not make any distinction between the vowel /i/ and the cor-
responding semivowel /j/, nor between the vowel /u/ and the semivowel /v/.
Today, we would say that this is phonemically correct, since the distinction be-
tween the vowel and semivowel is one of syllabicity, and it makes for a more
economical system to use the same character for both. The syllabic and asyllabic
sounds did not contrast in Old Icelandic, for example /i/ was only allowed in the
nucleus of the syllable, while /j/ was only allowed in the margin.
The reasoning behind the rising diphthong is not very clear, and will not be
discussed here. Suffice it to say that the FGT advocates the usage of ‹ e › in these
diphthongs, so that járn n. ‘iron’ should be spelt ‹ eárn › and jór m. ‘horse’, pre-
sumably, ‹ eór ›.

6. The letters of the language should be designed so that their shape reflects their
sound.

While the previous five points are uncontroversial, I would like to add a sixth
point, namely the first grammarian’s analysis of the vowel symbols. From a
modern point of view, the shape of a character in a script is arbitrary; the
historical development of the A from the head of an ox does not seem relevant to
14 E.C. Quiggin Memorial Lectures

the usage of this letter. The orthography of most languages seems to converge in
using this letter for an open vowel, and that is it. The first grammarian was
apparently of a different opinion. His background is the traditional threefold
distinction of a letter into its name (nomen), its shape (figura) and its sound (vox).
He accepted the five vowels ‹ a ›, ‹ e ›, ‹ i ›, ‹ o ›, ‹ u › of the Latin alphabet, and
incidentally of Ancient Nordic, as a natural basis. His aim was not to make a new
alphabet, but rather to add to it in a consistent manner. Now, after the rather
stable period of Ancient Nordic c. 200–500 AD, the processes of syncopation
and mutation led to the phonemisation of four new vowels in the language,
denoted in Old Norse orthography as ‹ y ›, ‹ ø ›, ‹ æ ›, ‹ ǫ ›. The first grammarian
obviously wanted these characters to be well motivated as far as their shape was
concerned, and his arguments are surprisingly to the point. As shown in Fig. 6,
the graphic traits of the proposed characters testify to their phonic status.16 His
arguments is as follows:

‹ y › has the stem of the ‹ i › and the branch of the ‹ u ›. Right! As Fig. 6 shows,
these are all high vowels, and ‹ y › is rounded like ‹ u › and fronted like ‹ i ›.

‹ ø › has the circle of the ‹ o › and the cross-bar of the ‹ e ›. Right! These are all
medium-high vowels, and ‹ ø › is rounded like ‹ o › and fronted like ‹ e ›.

‹ ę › has the shape of the ‹ e › and the loop of the ‹ a ›. Right! These are all non-
high vowels, and if we accept that the hook can be seen as a representation of
the ‹ a ›, ‹ ę › indeed has the shape of both. Alternatively, one can regard the
hook as a diacritical mark indicating a lower position of the vowel, and while
‹ e › is a medium-high front vowel, ‹ ę › is a low front vowel. Note that while
many of the early Old Icelandic manuscripts agree in using the letter form ‹ ę ›,
standard Old Norse orthography has the ligature ‹ æ › for this vowel.

‹ ǫ › has the circle of the ‹ o › and the loop of the ‹ a ›. Right! These are all back,
non-high vowels, and ‹ ǫ › has elements of ‹ o › as well as ‹ a ›. In many early
Old Icelandic (and Norwegian) manuscripts, a ligature of ‹ a › and ‹ o ›, namely
‹ ꜵ ›, was used for this sound. Once again, the hook can be regarded as a
16
The Korean hangul script has a similar basis in the phonic level, since characters in this
script reflect the actual pronunciation of the sounds. The hangul script was made from
scratch in the mid-fifteenth century under the reign of emperor Sejong the Great (1397–
1450 AD), and today it is the dominant script in Korea, although hanja characters are also
used as supplements. However, not even a Thor Heyerdahl would suggest that there was
any connection between the orthographical recommendations in FGT and hangul!
‘So that the writing may be less and quicker’ 15

diacritical mark for a lower position of the vowel, and ‹ ǫ › is indeed a low
vowel as opposed to the medium-high ‹ o ›.

On two points we are left in the dark. One is capitalisation. Manuscripts at the
time used majuscules as well as minuscules, although not always in a manner
which is consistent with modern usage. However, would the first grammarian
think that the distinction between minuscules and majuscules is actually a super-
fluous one? That was indeed the case for the German tradition of kleinschreibung,
i.e. using only minuscules (lower-case letters), as advocated by e.g. the gram-
marian Jakob Grimm, and made into a programme by the Bauhaus movement.
For this reconstruction, we will assume that the FGT would stick to klein-
schreibung, and thus simply do away with a distinction of minor value.17
The other point is punctuation. From other medieval traditions, we know of
elaborate systems of punctuation, using such marks as mid point, punctus elevatus,
punctus flexus, etc.18 However, the great majority of Old Norse manuscripts use
only a single punctuation mark, the dot. The most likely hypothesis is that the
FGT would find this mark sufficient, too.

Trying to reconstruct the orthography

The time has come for a reconstruction of the orthography. The three opening
lines of the FGT will be used as a test case (see the facsimile in Fig. 1 for com-
parison). The source is by necessity the text in W, the only preserved witness.
This has been rendered in diplomatic style, following the edition by Hreinn
Benediktsson (Reykjavík 1972, p. 206). In his edition, abbreviations are expanded
silently, but the rules for expansion have been given at the outset (p. 205).

17
Guðvarður Már Gunnlaugsson (personal communication) points out, however, that the
first grammarian may have wanted to use capitals at the beginning of sections, cf. p. 88, l.
21–22 in the manuscript (quoted from Hreinn Benediktsson’s edition, Reykjavík 1972, pp.
234–235, see also Einar Haugen’s edition, London 1972, pp. 24–27): ‘... allra helldz er ek
rít[ca] þa hofvð ſtafí ſtæʀi enn aðra í rítí er æigi ſtanda í verſ vpphafí ...’ (... in particular
since I [do not] write those capitals any larger than other (letters) in the text that do not
come at the beginning of a period ...). Note that Einar Haugen translates vers as ‘chapter’
rather than ‘period’. Also note that the addition of the negating particle -ca in rít[ca] is an
emendation of the text, originally introduced by Rasmus Rask. It was accepted by both
Einar Haugen and Hreinn Benediktsson in their editions.
18
The standard work in this respect is Malcolm B. Parkes, Pause and Effect, Aldershot:
Scolar Press, 1992.
16 E.C. Quiggin Memorial Lectures

Diplomatic transcription

[ J ] fleſtvm londvm ſetia menn a bækr annat tveggia þann froðleik er þar
innan landz | hefir giorz ęða þann annan er minniſamligaztr þikkir þo at
annarſ ſtaða[r hafi] helldr | giorz ęða lǫg ſín ſetia menn a bækr hverr þioð a
ſína tvngv

The first step is to normalise this text according to standard Old Norse ortho-
graphy, thus ironing out variation in the orthography of W. The process of
normalisation is well known and documented, although it should be pointed out
that there is some variation within the Old Norse standard.19 Normalising to
standard orthography requires, among other things, indicating vowel length by
accents and consonant length by gemination, introducing the older distinction
between ‹ œ › and ‹ æ › and between ‹ ø › and ‹ ǫ ›, using the vowels ‹ i ›, ‹ a › and ‹ u ›
in unstressed position, and using ‹ þ › in initial position and ‹ ð › in medial and
final position. This is how Einar Haugen did it in his edition and translation of
the work (London 1972, pp. 12–13):

Standard Old Icelandic orthography

Í flestum lǫndum setja menn á bœkr annat tveggja þann fróðleik, er þar innan-
lands hefir gǫrzk, eða þann annan, er minnisamligstr þykkir, þó at annars
sta[ðar hafi h]eldr gǫrzk, eða lǫg sín setja menn á bœkr, hver þjóð á sína tungu.

(In most countries men put into books either the great events that have come
to pass within their country, or whatever seems most memorable that has
occurred abroad, or men put their laws into books, each people in its own
language.)

19
For many, the orthography of the series Íslenzk fornrit (Reykjavík 1933–) is a point of
reference. It is based on the language of the early thirteenth century and thus makes a
distinction between /ø:/ and /æ:/ in the shape of ‹ œ › vs. ‹ æ ›, and between /ø/ and /ͻ/ in
the shape of ‹ ø › vs. ‹ ǫ ›. Some grammars and dictionaries use a slightly different notation.
For example, Ordbog over det norrøne prosasprog (Copenhagen 1989–) renders the first
opposition as ‹ ǿ › vs. ‹ ǽ ›. In the digital edition of the Eddic poems, Haraldur Bernharðs-
son has normalised the text according to the norm of the manuscript, GKS 2365 4° (c.
1270), which leads to a slightly younger orthography, in which /ø:/ and /æ:/ have been
merged in ‹ æ ›, and likewise /ø/ and /ͻ/ in ‹ ö ›. A specimen of Vǫluspá in GKS 2365 4°
(traditionally referred to as Codex Regius) is now available in the Menota text archive,
http://www.menota.org. Here, the text has been rendered on three levels: a facsimile
level, a diplomatic level and a normalised level (the latter according to the manuscript).
‘So that the writing may be less and quicker’ 17

The next step is, in comparison, a step into the unknown, i.e. into the complete
FGT orthography.20 Verner Dahlerup and Finnur Jónsson in their edition of the
first and second grammatical treatises made a similar reconstruction,21 but I have
taken this venture a step further with respect to the notation of nasality and also
provided a calligraphic version of the result. The opening of the FGT would then
look like this:

FGT orthography

í flęſtom lndom ſętıa mɴ á bǿcr. ȧɴat tuęɢıa þȧɴ fróþleic eſ þar iɴanlȧndſ
hęfer gǫrþſc. ęþa þȧɴ ȧɴan eſ miɴeſamlegſtr þyᴋer. þó at ȧɴarſ ſta[þar hafe
h]ęldr gǫrþſc ęþa lǫg ſn ſętıa mɴ á bǿcr. huęr þeóþ á ſna tuŋo.

For this reconstruction, I have established fourteen rules:22

1. A long consonant is rendered by a small capital, e.g. ‹ tveggja › > ‹ tveɢja ›

2. The dental fricative /þ/ is rendered as ‹ þ › in all positions, e.g. ‹ þjóð › >
‹ þjóþ ›
3. The short consonants /k/ and /s/ are rendered by ‹ c › and ‹ ſ ›, e.g. ‹ bœkr › >
‹ bœcr ›, ‹ setja › > ‹ ſetja ›
4. The cluster of the consonants /n/ and /g/ is rendered by ‹ ŋ ›, e.g. ‹ tungu › >
‹ tuŋu ›
5. The cluster of a dental consonant, i.e. /þ/, /d/ or /t/, and /s/ is not rendered
by ‹ z ›, but by its separate components, e.g. ‹ gǫrzk › > ‹ gǫrþſc ›
6. A short, oral vowel has no diacritical mark, not even ‹ i ›, e.g. ‹ hefir › > ‹ hefır ›
7. A short, nasal vowel is indicated by a superscript dot unless it is unstressed,
e.g. ‹ lǫndom › > ‹ lndom ›
8. A long, oral vowel is indicated by an acute accent, e.g. ‹ bœcr › > ‹ bǿcr ›
9. A long, nasal vowel is indicated by the combination of a superscript dot and
an acute accent, e.g. ‹ sína > ‹ sna ›

20
This requires a font with the necessary characters. I have used Andron Corpus, developed
by Andreas Stötzner in Leipzig, cf. http://www.signographie.de.
21
See their edition, Den første og anden grammatiske afhandling i Snorres Edda, København
1886, pp. 19–49.
22
Note that in the examples, I have only made the change specified in each rule; thus
‹ tveggja › > ‹ tveɢja › is the result of rule (1), while ‹ tveɢja › > ‹ tueɢıa › is the result of rule
(10). The cumulative effect of the rules is that some words change shape radically, e.g.
‹ tveggja › > ‹ tuęɢıa ›.
18 E.C. Quiggin Memorial Lectures

huat þá ſcal at hafa. cuaþ ec. huat þá nėma ſna hȯnom ſuá ſcýr dme þęıra
gręina. eſ hȧɴ ſcılr ŋuar áþr ueſa. at þá þyᴋeſc hȧɴ of ſęiɴ uerþa tıl at mǽla
ſıalfr á mót ſér. oc uerþa fyʀe at bragþe. ėn þęır eſ ęʟa mẏnde fífla hȧɴ oc caʟa.
ſėm uǽre ſpacara. ef þęgþe. nú ero hér þau dme eſ bráþa fȧŋſ fndoſc þęır. ėn
ſíþan nǫᴋoro leóſlegar tıl málſ fǿrþ oc ſcılniŋar. ú bé uʙe. ſęcr ſęᴋr. h dó
hǫᴅo. áfarar aꜰarar. þagat þaɢat. ǫl ǫʟ. frȧme frȧᴍe. uina uiɴa. crapa craᴘa.
huęr huęʀ. fúſ fúꜱ. ſceót ſceóᴛ. ú bé. þat ero tuau nǫfn tuęɢıa bóc ſtafa. ėn uʙe
þat eſ ęinſ mȧɴſ ęıᴛ nafn. ſęcr eſ ſcógar maþr. ėn ſęᴋr eſ ílát. h dó þá eſ hǫlga
troʟ dó. ėn hęyrþe tıl hǫᴅo þá eſ þóʀ bar hueʀeɴ. [...] bętra eſ huęrıom fyʀ
þagat. ėn ȧɴaʀ hafe þaɢat. ęıgı ero ǫl ǫʟ at ęino. męıre þyᴋer ſtýre mȧɴ ſnſ
frȧme. ėn þeꜱ eſ þılıornar byɢuer frȧᴍe. ſá eſ meſtr guþſ uina. eſ meſt uıʟ tıl
uiɴa. uaþa opt tıl cırcıo crapa. þó at þar fáe lęıþ craᴘa. huęr cȯna oc huęʀ carl
maþr ſcyldo þeſ fúſ. ſėm guþ eſ fúꜱ. þá mno þau tıl góþra uerca ſceót. oc hafa
guþſ hyʟe ſceóᴛ. nú m þȧɴ mȧɴ. eſ ríta uıʟ eþa nėma at uáro mále rıtet. ȧɴat
tuęɢıa helgar þýþiŋar eþa lǫg eþa áᴛ uíſe eþa ſuá huęrege eſ maþr uıʟ ſcẏn
ſȧmlega nyt ſme á bóc nėma eþa cɴa. nda ſé hȧɴ ſuá lítel látr í fróþ lęıx
áſteɴe at hȧɴ uıle nėma lítla ſcẏn ſme hęldr ėn ŋua. þá eſ á meþal uerþr iɴar
męıre. þá leſe hȧɴ þeᴛa cápítúlom uȧndlega. oc bǿte ſėm í mǫrgom ſtǫþom
mn þurfa. oc mete uıþ lęıtne mna ėn uárcẏɴe ócnſco. hafe ſtaf róf þeᴛa eſ
hér eſ áþr rıtet. nz hȧɴ fǽʀ þat. eſ hȯnom lícar bętr.

aáȧ ǫ eéė ę ıíi oóȯ øǿ uú yýẏ

bʙ cᴋ dᴅ fꜰ gɢ ŋh lʟ mᴍ nɴ pᴘ rʀ ſꜱ tᴛ xþ ̄ ͛

Text 1. The final part of the FGT in the reconstruction of its own recommended orthography,
made according to the rules established in this article. Note that the explanation of ‘áfarar
aꜰarar’ is lacking in the manuscript. It is an open question whether this goes back to the
original or it was lost during transmission

10. The semivowels /j/ and /v/ are rendered by ‹ i › and ‹ u ›, e.g. ‹ tveɢja › >
‹ tueɢıa ›, ‹ hver › > ‹ huer ›
11. The rising diphthongs /ja:/, /jo:/ and /ju:/ are rendered by ‹ eá ›, ‹ eó › and
‹ eú ›, e.g. ‹ þjóð › > ‹ þeóþ ›
12. The vowels /æ/ and /æ:/ are rendered by ‹ ę › and ‹  ›, e.g. ‹ huer › > ‹ huęr ›
(where ‹ e › reflects /æ/, the i-mutation of /a/) and ‹ mæla › > ‹ mla › (not in
this text)

Finally, I have added two phonologically-based rules:

‘So that the writing may be less and quicker’ 19

What should one do, I said – except to show him such clear examples of the
distinctions he did not notice before that he will hurry and take back his own
words, so that he may get ahead of those who otherwise would mock him and
say he would have been wiser to have kept his mouth shut. Here are the
examples that could be found in a hurry, which will then be arranged in
sentences for clarity’s sake: ú bé, uʙe; secr, seᴋr; h dó, hǫᴅo; áfarar, aꜰarar;
þagat, þaɢat; ǫl, ǫʟ; frame, fraᴍe; uina, uiɴa; crapa, craᴘa; huer, hueʀ; fús, fúꜱ;
sceót, sceóᴛ. Ú bé are the names of two letters (u b), but Ubbi (uʙe) is a man’s
name. A forest-dweller is outlawed (secr), but a sack (seᴋr) is a bag. A tall
woman died (h dó) when Hǫlgatroll died; but you could hear the handle
(hǫᴅo) when Thor carried the kettle. [...] It is better for every man to have
been silent (þagat), rather than let another have silenced (þaɢat) him. Not all
(ǫʟ) beers (ǫl) are alike. The skipper’s fame (frame) is held to be greater than
his who bunks on the forward (fraᴍe) deck. He is the greatest of God’s friends
(uina) who will work (uiɴa) the hardest for Him. People often wade to church
through wet snow (crapa), even though it makes a difficult (craᴘa) path. Every
(huer) woman and every (hueʀ) man should be desirous (fús) of that of which
God is desirous (fúꜱ). Then they will be quick (sceót) to do good deeds and
quickly (sceóᴛ) gain the grace of God. Now any man who wishes to write or to
learn that which is written in our language, whether it be sacred writings or
laws or genealogies or whatever useful knowledge a man would learn or teach
from books if he is humble enough in his love of learning so that he will
rather gain a little insight than none, until there is a chance for more – then let
him read this treatise with care, and improve it, as it no doubt needs in many
places, let him value my efforts and excuse my ignorance, and let him use the
alphabet which has already been written here, until he gets one that he likes
better:

aȧ ǫ eė ⟨ę⟩ ıi oȯ ø u yẏ bʙ cᴋ dᴅ fꜰ gɢ ɡh lʟ mᴍ nɴ
pᴘ rʀ sꜱ tᴛ xþ ̄ ͛

Text 2. The translation by Einar Haugen of the section which corresponds to the facing page,
from his edition of the FGT, pp. 31–33. The pair ⟨ ę  ⟩ was evidently left out by mistake.

13. The unstressed vowels are rendered by ‹ e ›, ‹ a › and ‹ o ›, not by ‹ ı ›, ‹ a › and

‹ u ›, e.g. ‹ hafi › > ‹ hafe › and ‹ tungu › > ‹ tuŋo ›
14. According to Verner’s law, the relative particle ‹ er › is rendered as ‹ eſ ›, and
the verb ‹ vera › as ‹ ueſa ›.
20 E.C. Quiggin Memorial Lectures

Fig. 9. The last page of the First Grammatical Treatise in the orthography reconstructed in this
article and penned by Bas Vlam (2011), based on the script of the first 14 lines of Reykja-
holtsmáldagi (c. 1130–1150). This section is identical to the one in texts 1 and 2 on the pre-
ceding pages. Note that the small capital ‹ ᴋ › was rendered in a kappa-like form in later Ice-
landic script, ‹  ›, but the standard Carolingian shape has been chosen here.

The recommended alphabet at the very end of the FGT is a particular chal-
lenge, since it is not in the orthography of Codex Wormianus, nor is it a likely
version of an alphabet in the original. The explanation may be that the original
‘So that the writing may be less and quicker’ 21

Fig. 10. The last page of the First Grammatical Treatise as preserved in AM 242 fol, p. 90 (=
fol. 45v), l. 12–32. The reconstruction in Fig. 9 on the facing page begins with ‘huað’ in the
middle of the first line. Note the difference in style and orthography – and in the rendering of
the recommended alphabet in the last couple of lines.

simply did not conclude with any alphabet, but when the text was copied by later
scribes who understood that the language had changed, someone along the line of
copying may have tried to spell out what the alphabet would have looked like.

Concluding thoughts

Even if there is some variation in the few preserved specimens of early Icelandic
Carolingian script, the reconstruction presented in this article, based on the
script in Reykjaholtsmáldagi, has chosen what is probably the earliest specimen as
its model. Reykjaholtsmáldagi is contemporaneous with or possibly somewhat
22 E.C. Quiggin Memorial Lectures

earlier than the presumed original of the FGT. As mentioned above, there is only
a handful of Icelandic specimens which can unequivocally be said to be in Caro-
lingian style, perhaps no more than three specimens, and of these only one full-
length manuscript, AM 674 a 4°.
The reconstruction in this article has been carried out without abbreviations
as it seems that the early manuscripts used abbreviations somewhat sparingly,
and especially Reykjaholtsmáldagi. If abbreviations, e.g. occasional nasal strokes,
had been introduced in the reconstruction, the overall look would not have been
much different. It should also be added that the original manuscript of the FGT
probably would have been written in a single column, as is the case with other
early Icelandic manuscripts (except for AM 237 a fol, which was written in two
columns). One might also surmise that it would have been written in a rather
simple style, with little or no usage of initials and litterae notabiliores. These traits
seem to have been introduced somewhat later, perhaps around 1200. So, for what
it is worth, the palaeographical reconstruction presented here may give a fair
indication of what the original might have looked like.
The orthography is another matter. What did surprise me when trying to
follow all the recommendations of the FGT was how different the orthography
turned out to be. There is a considerable distance from it to the standard norma-
lised orthography of Old Icelandic (and Old Norwegian), even if this is based on
the language of the early thirteenth century. There is also considerable distance
from it to any of the actual manuscripts of the time, as they can be accessed in
collections such as the aforementioned Early Icelandic Script by Hreinn Bene-
diktsson (1965) and Sýnisbók íslenskrar skriftar by Guðvarður Már Gunnlaugsson
(2007). Even if the majority of manuscripts of this period have been lost, it
seems likely that even for scribes of the time, the orthography presented here
would have looked strange and possibly difficult to follow. For this reason, one
might think that the first grammarian would have chosen a less deviant ortho-
graphy when presenting his recommendations.
We cannot know exactly what this orthography would have looked like, since
there is some variation in the preserved specimens, and there would surely have
been more variation in the now lost manuscripts. However, since the fourteen
first lines of Reykjaholtsmáldagi have the distinction of being contemporaneous
with or possibly antedating the FGT, they are our best guide to the prevalent
orthography at the time of the FGT. To the extent that they can provide con-
clusive answers, we see that they make no distinction between short and long
vowels, they make no indication of nasal vowels, and long consonants are only
occasionally rendered by gemination, never by small capitals. So if we assume
‘So that the writing may be less and quicker’ 23

that the first grammarian would have chosen a less deviant orthography, the one
in Reykjaholtsmáldagi might be the best on offer. However, the initial fourteen
lines of Reykjaholtsmáldagi leave too many questions unanswered. So perhaps the
final answer to my second question is that we really cannot know, but that our
best guess is that the first grammarian would have been hesitant to follow his
own recommendations if he were to have put them into practice.

I would like to thank Judy Quinn and her colleagues at the Department of Anglo-
Saxon, Norse and Celtic for the invitation to give the E.C. Quiggin Memorial Lecture
in 2011 and for their hospitality when I was in Cambridge. I would also like to thank
Guðvarður Már Gunnlaugsson at Stofnun Árna Magnússonar í íslenskum fræðum in
Reykjavík for a number of pertinent and helpful comments to this article.
The Department of Anglo-Saxon, Norse and Celtic offers programmes of study, at
both undergraduate and graduate level, on the pre-Norman culture of the British
Isles in its various aspects: historical, literary, linguistic, palaeographical,
archaeological. The Department also serves as a focal point for scholars visiting
Cambridge from various parts of the world, who are attracted to Cambridge by the
University Library (one of the largest in the world), the collections of Anglo-Saxon
and Celtic manuscripts in the University and various college libraries, the
collection of Anglo-Saxon, Celtic and Scandinavian coins in the Fitzwilliam
Museum, or the rich collection of Anglo-Saxon artefacts in the Museum of
Archaeology and Anthropology. It is possible for the Department to host a small
number of Visiting Scholars each year.

Information on any aspect of the Department’s activities can be obtained by writing

to: The Head of Department, Department of Anglo Saxon, Norse and Celtic, Faculty
of English, 9 West Road, Cambridge, CB3 9DP.

Further information on the Department, on the Anglo-Saxon, Norse and Celtic

Tripos, and on opportunities for postgraduate study, is available on our website:
www.asnc.cam.ac.uk.
E. C. QUIGGIN MEMORIAL LECTURES ISSN 1353-5722

Sources of Mediaeval Gaelic History

1 John Carey, The Irish National Origin-legend: Synthetic Pseudohistory (1994)

ISBN 0 9517339 8 2

2 Dauvit Broun, The Charters of Gaelic Scotland and Ireland in the Early and Central Middle Ages (1995)
ISBN 0 9517339 9 0

3 David N. Dumville, Councils and Synods of the Gaelic Early and Central Middle Ages (1997)
ISBN 0 9532172 0 5

4 T. M. Charles-Edwards, The Early Mediaeval Gaelic Lawyer (1999)

ISBN 0 9532172 1 3

5 John Hines, Old-Norse Sources for Gaelic History (2002)

ISBN 0 9543186 3 3

6 Pádraig P. Ó Néill, Biblical Study and Mediaeval Gaelic History (2003)

ISBN 1 904708 00 5

7 Tadhg O’Keeffe, The Gaelic Peoples and their Archaeological Identities, A.D. 1000–1650 (2004)
ISBN 1 904708 10 2

8 Dáibhí Ó Cróinín, The Kings Depart: The Prosopography of Anglo-Saxon Royal Exile in the Sixth and Seventh
Centuries (2007)
ISBN 0-9554568-2-7 / 978-0-9554568-2-4

Quiggin Memorial Lectures

9 Erich Poppe, Of Cycles and Other Critical Matters. Some Issues in Medieval Irish
Literary History and Criticism (2008)
ISBN 978-0-9554568-5-5

10 Henrik Williams, Rune-stone Inscriptions and Queer Theory (2008) ISBN 978-0-9554568-7-9

11 Uáitéar Mac Gearailt, On the Date of the Middle Irish Recension II Táin Bó Cúailnge
(2010)
ISBN 978-0-9562353-2-9

12 Carole Hough, Toponymicon and Lexicon in North-West Europe: ‘Ever-Changing

Connection’ (2010)
ISBN 978-0-9562353-3-6

13 Liam Breatnach , The Early Law Texts Senchas Már and the Question of its date
(2011) ISBN 978-0-9562353-9-8

Copies of these lectures may be obtained from the Departmental Secretary, Department of Anglo-Saxon,
Norse and Celtic, Faculty of English, 9 West Road, Cambridge, CB3 9DP; telephone 01223–335197.

For a complete list of all available publications, please see our website: www.asnc.cam.ac.uk