Estonian corpus
===============

(1) Character coding info

We have used extended ASCII for coding of some especial 
Estonian characters. They are:
	a Umlaut - code 132,
	A Umlaut - code 142,
	o Umlaut - code 148,
	O Umlaut - code 153,
	u Umlaut - code 129,
	U Umlaut - code 154 (all as in German);
	a "o" with a "^" over it - code 147.
Then

	"Õ" means a capital "O" with a "~" over it, 
	"&z;" means a "z" with a "v" over it, 
	"&Z;" means a capital "Z" with a"v" over it, 
	"&s;" means a "s" with a "v" over it, 
	"&S;" means a capital "S" with a "v" over it, 
	"&c;" means a "c" with a "v" over it, 
	"&C;" means a capital "C" with a "v" over it. 

	"&t;" occurs once (originally t^) - it may be an error

(2) Documentation suipplied by Turku

                      MANUAL OF ENCODING                     
                         TO ACCOMPANY                        
                  THE TARTU COMPUTER CORPUS                  
                     OF THE CONTEMPORARY                     
                      ESTONIAN LANGUAGE                      


                 Translated from Estonian by                 
                        Madis Saluveer                       


                       Tartu April 1992                      


The following pages present the coding manual used during the 
editing and computer mark-up  of  the  Estonian  texts.  This 
manual is based on the "Manual of Information"  for  the  LOB 
corpus  by  S.  Johansson,  G. Leech and H. Goodluck and  was    
updated  according  to  the requirements of the Text Encoding 
Initiative TEI P1 Draft Version 1.0.

Coding manual: Table of contents

1  Organisation of the material
2  Coding key
3  Textual material included/excluded
4  Spacing and ordering
5  Typographic shifts
6  Headings
7  Symbols denoting paragraphs and lines
8  Sentence mark-up
9  Direct speech
10 Quotations
11 Capitalisation and proper names
12 Lists
13 Foreign language material
14 Other non-contemporary material
15 Abbreviations
16 Hyphen and dash
17 Mathematical expressions
18 Typographical errors


1 Organisation of the material

1.1. The corpus starts with the marker for the first text and 
ends with the end-of-corpus symbol.

1.2. The length of a text  sample is 2,000 + 200  words.  The 
sample text must be  a  comparatively  integral whole. A word 
is defined ortographically  as  a  character  or sequence  of 
characters  surrounded  by  blank  spaces  (and including no 
blank spaces).

1.3. Each corpus text starts with a tag giving  its  absolute 
number and category number and  ends  with  the  end  of-text 
symbol and a figure giving the number of words in the text.

1.4. Each corpus text is  preceded  by  a  standardised  data 
description. The data description is marked by the tag  <file 
description> which is organised as follows:

<file description>
   <title statement>
    <title>      </title>
      <sponsor>         </sponsor>
      <funding agency>              </funding agency>
      <principal   reseacher>          </principal reseacher>
    <statement of responsibility>
          <name>                </name>
          <role>                </role>
    </statement responsibility>
   </title statement>
<edition statement>
 <edition>         <date>       </date>            </edition>
 <statement of responsibility>
     <name>                     </name>
     <role>                     </role>
</edition statement>
<publication statement>
  <publication>
      <publisher>                </publisher>
      <place>                    </place>
      <date>                     </date>
  </publication>
</publication statement>
<series statement>
  <series title>                    </series title>
  <statement of responsibility>
        <name>                    </name>
        <role>                    </role>
  </statement of responsibility>
</series statement>
<source description>
</source description>
</file description>
<encoding declaratious>
</encoding declaratious>
<revision history>
</revision history>
<text>
</text>
<TEI.1>

1.5. Only those blank spaces in the  file  description  which 
can be supplied for a given text are  filled  in.  The  other 
remain unfilled.

1.6. Some data are permanent, ie the same for all texts. They 
include:

2. Coding key

2.1. Alphanumeric characters represent themselves  (including 
o, a, o, u)

A = A         B = B          C = C           1 = 1
a = a         b = b          c = c           2 = 2, etc.

except for

   = oe                  = ae                 = ss
z  = z^                s = s^               O = O^

Some other foreign letters untypical for  Estonian are  marked 
the same way, eg   r = r^,   e ,   etc.

2.2. Other symbols

In general, punctuation marks and other symbols used for  text 
delimitation represent themselves:

, = ,      . = .     ? = ?     / = /    ( = (    etc.

It should be noted:

- = hyphen or minus is encoded without blank spaces
- = dash is preceded and followed by a blank space
" " = the begin and end  quotation  marks  are  excluded  and 
      substituted by the markers <q> and </q>

2.3. As regards numerical expressions, see 17.4.

2.4. Words are not broken at the end of a line.

2.5. All changes in texts are  marked  by  English  words  or 
letters included within the symbol < > where < > denotes  the 
beginning of the change and </ > denotes its end:

<p>  = begin new paragraph or line
</p> = end of paragraph
<s>  = begin sentence
</s> = end of sentence, etc.

2.6. The coding markers may be simple (consisting of only one 
main symbol) or compound (the attribute  following  the  main 
symbol represents an aspect or value  of  the  main  symbol). 
Attributes and their values may vary and will be presented in 
the corresponding subparts of the Manual.
Attributes are separated from their main symbol  by  a  blank 
space. The value of an attribute is  presented  after  the  = 
symbol without a blank space: <note source=ed>.

3 Textual material included/excluded

3.1. Sentences used as  "tantalizers"  at  the  beginning  of 
prose texts are included. They are preceded  by  a  paragraph 
marker <p> and followed by the end-of-paragraph marker  </p>. 
The same applies to "story so far" summaries.

3.2. Commentaries and notes  extraneous  to  the  source  are 
omitted and represented by a comment tag <note>.

3.3. The marker <note> has the following attributes:
1) n denoting the number of the comment, eg
     <note n=24>
2) type denoting the type of the comment,eg
     <note n=1 type=explanation>
3) source identifying the author of the comment, eg
     <note source=ed(itor)>
or               =comp(iler)>
or               =transcriber>,  etc.
4) place indicating the locality of the comment in the  text, 
eg
<note place=foot>
or         =end>
or         =inline>
or         =app(endix)1>
All the attributes will be encoded in the above order.

3.4. Long foreign quotations are excluded and represented  by 
the tag <quot> in their original place. Foreign quotations up 
to 30 words are included. Even longer foreign quotations  may 
be  included  if  their  exclusion  will   seriously   hamper 
textunderstanding 
The marker <quot> has the attribute "foreign" which  identifies 
the language of the quotation, eg
     <quot foreign=Fre>

3.5. Long poetry quotations are excluded and replaced by  the 
marker <quot poem> in the original place in the text.

3.6. As regards mathematical expressions see 17.3.

4 Spacing and ordering

4.a. Spacing

4.1. A single space indicates a typographic word-boundary  in 
the source text. A space may not occur within a word.

4.2. / occurring between words is followed by a space, except 
in the case of and/or (ja/voi) which is  coded  as  a  single 
word. Np space is inserted after / in numerical expressions.

4.3. A space follows the punctuation marks . , ; : ? ! as  in 
printed texts. These marks are not preceded by a space.

4.4. A dash is precede and followed by a single space  as  in 
printed texts. A hyphen is neither preceded nor followed by a 
space.

4.5.An opening bracket ( is preceded but  not  followed by  a 
space. A closing bracket ) is followed but not preceded  by a 
space.

4.6. A space is not used between markers and units denoted by 
them. A space separates markers and units not related to each 
other. No space occurs between two markers denoting the  same 
unit.

4.b. Ordering

4.7. If several markers are applied to one and the same  word 
or phrase at the same point in the text then the order of the 
opening markers is as follows:
1) the typographical shift marker <highlighted> occurs before 
the proper name marker <propname>,
2) the typographical shift marker <highlighted> precedes  the 
sentence marker <s>,
3)  the  proper   name   marker   <propname>   precedes   the 
abbreviation marker <abbrev>,

<note resp=eci> All <abbrev> elements have been replaced by <abbr> to
be conformant with TEI 2 DTD </note>

4) the sentence  marker  <s>  precedes  the  quotation  marks 
marker <q>,
5) the quotation marks  marker  <q>  precedes  the  quotation 
marker <quot>. For example,
<propname type=institution><abbr>USA</abbr></propname>

4.8. If several markers are applied to one and the same  word 
or phrase at the same point in the text then the ordering  of 
opening and closing  markers  must  be  symmetrical,  ie  the 
leftmost opening marker must be the  same  as  the  rightmost 
closing marker. In other words,  the  order  of  opening  and 
closing markers is reversed, eg
<q><highlighted                    rendition=italic><propname 
type=person>Ivanov</propname></highlighted></q>

4.9. If several markers occur at the same point in  the  text 
and their closing order is not the same then they  should  be 
placed so that the markers denoting longer units will  remain 
outside the markers denoting shorter units, eg
<highlighted  rendition=bold><s>See  kohta  palusime  raakida 
<propname                                   type=person>Harri 
Aguraiujal</propname></s></highlighted>
<s>Tanavusel ERKI moeshow'l oli olulisel kohal
<p>
     ABSURD.
</p></s>

5 Typographic shifts

5.1. The beginning of a typographic  shift  in  the  text  is 
denoted  by  the  marker  <highlighted>  and   its   end   by 
</highlighted>.

5.2. The shift type is indicated with the help  of  attribute 
"rendition". "Rendition" may have the value: italic,  spaced, 
sans, bold, underscored, etc., eg
<highlighted rendition=italic>

5.3. The symbol of typographic shift may occur within a word, 
eg S<highlighted rendition=bold>ona</highlighted> anti...

5.4. The symbols . , : ? ! and other  punctuation  marks  are 
regarded  as  neutral  between  different  shift  types.  The 
end-of-sentence   punctuation   mark   occurs   before    the 
typographic shift end marker when the whole  sentence  is  in 
the corresponding shift.

6 Headings

6.1. The headline is preceded by the  marker  <headline>  and 
followed by the marker </headline>. The markers are placed on 
a separate line.

6.2. In the source text headings are sometimes   placed   not 
at the beginning of a text but somewhere inside the text.  In 
corpus they are appropriately repositioned at  the  beginning 
of the text.

6.4. In the original headings may be characterised by special 
typographic  features  (only  capital   letters,   figurative 
shifts, etc.). Such headlines will be unified  to  the  first 
capital letter and the remainder be printed in small letters.

6.5. Other editorial headings (eg "continued on p. 3",  "Next 
week:...", etc.) are excluded.

6.6. The occurrence of "Editor" or the name of the author  at 
the end of the article is coded with the heading markers.

7 Symbols denoting partagraphs and lines

7.1. The beginning of a new paragraph  is  indicated  by  the 
symbol <p> placed on a separate line  before  the  paragraph. 
The end of a paragraph is indicated by the symbol </p> placed 
on a separate line.

7.2. A paragraph is always indented.

7.3. The paragraph marker is also used:
     a) to denote the lines of a poem
     b) to denote breaks in the  text,  such  as  a  two-line 
        space, asterisks, etc.
     c) to mark sub-headlines
     d) to indicate "tantalisers" at the beginning or  within 
        a text(eg words indicated by  a  typographical  shift 
        in a newspaper, etc.).

8 Sentence mark-up

8.1. The beginning of a sentence is indicated by  the  marker 
<s>, the end - by </s>. Both markers are placed on  the  same 
line with the sentence.

8.2. The marker indicating the beginning of a sentence is not 
used at the beginning of a headline. If a heading consists of 
more than one sentence than the second sentence is marked  by 
<s> at the beginning and </s> at the end.

8.3. Sentences are separated by the punctuation marks .  ?  ! 
... . A semicolon is  never  considered  a  punctuation  mark 
separating sentences.

9 Direct speech

9.1. The markers <q> and </q> are  used  to  indicate  direct 
speech.  The  quotation  marks  used  in  the  original   are 
excluded.

9.2. When the reporting clause  precedes  or  follows  direct 
quotations/direct speech both of them are marked by  separate 
sentence markers. In addition, the reporting clause  and  the 
direct speech together are marked together as one sentence:
<s><s>Isa  kusis  akki</s>:<s><q>Kas  tahate   homme   kalale 
tulla?</q></s></s>
<s><s><q>Sundis poiss,</q></s><s>lausus Andres.</s></s>

9.3. When the reporting clause is placed  in  the  middle  of 
direct speech consisting of one sentence, the whole is marked 
with  sentence  markers  and  direct  speech   markers.   The 
reporting clause is separated from the rest of  the  sentence 
with markers <in quot> and </in quot>
<s><s><q>Aga,  <in  quot>huudis  Maie,  </in  quot>kes   siis 
niiviisi tohib teha!</q></s></s>

9.4.  When  the  introductory  verb  is  placed  between  two 
sentences in direct speech then all the three  form  separate 
sentences marked by the sentence  markers.  In  addition  the 
whole is marked with the sentence marker:
<s><s><q>Oh!</q></s>  <s>huudis   Maie:</s>   <s><q>Kull   on 
tore!</q></s></s>

9.5. An included quotation is separated from  the  preceeding 
and following reporting clause with sentence markers and  the 
reporting  clause  is  considered  as  one   sentence   whose 
end-of-sentence marker comes at the end of the second half of 
the reporting clause:
<s><s>Kui   tudruk   utles:<s><q>Seda   poleks   ehk   tarvis 
olnud;</q></s>vastas     Jaan     kindlalt:</s>     <s><q>Oli 
tarvis</q></s></s>

9.6. When the name of the speaker is  not  mentioned  in  the 
reporting clause then the speaker is marked as follows:
<s><s>Kui   tudrukutles:<s><q>Seda    poleks    ehk    tarvis 
olnud;</q></s>vastas       Jaan        kindlalt:</s><s><q>Oli 
tarvis.</q></s></s>

9.7. In two special cases the presence of quotation marks  is 
denoted but their essence has to be specified by  the  corpus 
user in the future:
a) words or phrases of the meta-level
<s>Lause  <q>Tana  on  ilus   ilm!</q>   on   oeldistaiteline 
olemislause.</s>
<s>Sona <q>staadion</q> kaandub nagu <q>seminar</q>.</s>
b) irony, transfer of meaning, etc.:
<s>Tuli aeg, mil mulle suruti pahe <q>kuninga kubar</q>,  mil 
pidin teenima Vene tsaari.</s>

10. Quotations

10.1 Text  excerpts  of another author directly indicated as 
such are considered as quotations.

10.2. A quotation is marked with <quot> and its beginning and 
</quot> at its end.
<s><s>Filoloogiadoktor     Aarne     Vinkel     kirjutab:</s> 
<quot><q><s>Liivi luulekeele rahvalikkus avaldub  nii  sellle 
sonavaras,    lauseehituses    kui    ka    varsi     uldises 
intonatsioonis.</s> <s>Kohati arendab  Liiv  stroofide  kaupa 
rahvalaulu-laadilist varsket niidistikku.</s></q></quot></s>

10.3. When a quotation begins in the middle of a sentence  or 
forms  an  inseparable  part  of  the  reporting  clause   no 
sentence marker is used within the quotation:
<s>Fr. Tuglas on selle kohta oelnud,  et  <quot><q>koik  meie 
paremad   kirjanikud   ...   on   ikka   midagi    pooldanud, 
idealiseerinud voi polanud.</q></quot></s>

10.4. Whwn quotations extend  over  two  or  more  paragraphs 
there are opening quotation marks at the  beginning  of  each 
paragraph but closing quotation marks only at the very end of 
the quoted section. In such cases  the  marker  <q>  is  used 
at the absolute beginning of the quotation and  </q>  at  its 
absolute end.

10.5. Parts of a quotation or  dialogue  containing  markedly 
deviant usage from standard language are denoted  by  special 
markers (see 13-14).

11 Capitalisation and Proper Names

11.1. Continuing capitalisation is not indicated  separately, 
eg
<s>USA HOIATAB NSV LIITU...</s>

11.2. The marker <propname> is used to denote proper names.It 
may have a specifying attribute "type".

11.3. The attribute "type" may  have  values  person,  place, 
institution, product, event:
<s><propname    type=person>Jaak    Mand</propname>    tootab 
<propname           type=place>Tallinnas            <propname 
type=institution>TTUs.</propname></s>

11.4. When the user cannot specify the value of the attribute 
"type" it is adviseable to use the marker <propname> only.

11.5. The markers coding proper names should cover  a  proper 
name as widely as possible,  and  pay  no  attention  to  its 
definition:
<propname   type=institution>koondise   "Marat"   2.   tsehhi 
ametiuhingukomitee</propname> esimees

11.6. In case of contracted constructions where one main word 
has  several  attributes  the  whole   construction  should be 
inserted  between common markers, eg:
<propname   type=institution>Tartu   1.,   2.,   3.   ja   5. 
Keskkool</propname>

11.7. When a propper name has been turned into a  lower  case 
attribute it will not be marked as a proper name:
Toi talle brusseli kapsast.

11.8. When a lower case foreign word such as de or von occurs 
as part of a name it will be enclosed within the proper  name 
markers:
President <propname type=person>de Gaulle</propname>

12 Lists

12.1. The beginning of a list is marked with the marker <list>, 
the  end - </list> which are placed on a separate line.

12.2. When a list is  ordered  (numerically,  alphabetically) 
the marker will have an attribute "type=ordered".

12.3. Every item of a list will be  preceded  by  the  marker 
<item> and followed by the marker </item>  and  placed  on  a 
separate line. The dash and numbers at  the  beginning  of  a 
list line will be included between markers. For example,  the 
list
1) 2 muna
2)1 liiter piima
will be marked as follows:
<list type=ordered>
<item>1)2 muna</item>
<item>2)1 liiter piima</item>
</list>

13 Foreign language material

13.1. Words,  expressions,  sentences  and  paragraphs  in  a 
foreign language will be marked only when they are  indicated 
by means of italics, boldface, capitalisation  or  any  other 
typographic means.

13.2. The foreign words and expressions will be  preceded  by 
the marker <foreign> and followed by the marker </foreign>.

13.3. When possible the  language  of  the  foreign  word  is 
indicated with the help of an attribute "lang=", eg
<foreign lang=Eng>
<foreign lang=Rus>
The attribute has 3 digits to indicate the  language  by  its 
English name.

13.4. Foreign titles of books, operas, etc are marked  unless 
they consist simply of a name:
<foreign lang=Ita>La Traviata</foreign>, but: Fidelio

13.5. As regards foreign abbreviations, see 15.5.

14 Other non-contemporary material

14.1. Material deviating from  modern  literary  Estonian  is 
marked with the markers <nonst> and </nonst>

14.2. This marker may  have  an  attribute  "type"  with  the 
values old, spont(aneous), dial(ectal), slang, etc.

14.3. Parts of a dialogue containing occasional  non-standard 
characteristics (double  negation,  wrong  government,  etc.) 
will not be marked with this marker.

14.4. Quotations  from  sources  published  before  1918  are 
marked with the attribute "type=old".

14.5.     Dialectal     usage      not      mentioned      in 
"Oigekeelsussonaraamat" (Dictionary of   ) but  mentioned  in 
"Murdesonastik" (Dictionay of Dialects) are marked  with  the 
symbol <dial>:
<nonst type=dial>tetta</nonst>

14.6. Short forms indicated by means of an apostrophy are not 
marked, eg ma'i (=ma ei).

14.7. A sentence or  paragraph  containing  the  features  of 
spontaneous oral presentation will be marked as  a  whole  by 
means of the attribute "spont(aneous)", eg:
Tema     ema pahandas:<q><nonst      type=spont>Mina       ei 
taa!</nonst></q>

14.8. Slang will be marked with the symbol <type=slang>.

14.9.  Neologisms,  nonce-forms,  broken  Estonian,  etc.  is 
denoted with the symbol <nonst>, eg
<nonst>slaagivad</nonst>=(laulavad)

15. Abbreviations

15.1.  Abbreviations  are  coded  by  means  of  the  markers 
<abbr> and </abbr>

15.2. The  marker  <abbr>  may  have  an  attribute  "full" 

<note resp=eci>This attribute has been changed to "expan" for TEI2
conformance </note>

representing  the  full  form  of  a  rare  or  difficult  to 
understand abbreviation/acronym.  The full form is  presented 
in the nominative case:
<s>Jaan          on          <abbr           expan="Tallinna 
Tehnikaulikool">TTU</abbr> tudeng.<s>

15.3. Abbreviations are coded whether or not they end  in  an 
abbreviation point. An abbreviation point will be followed by 
a closing marker. A case ending at the end of an abbreviation 
will be followed by the closing marker:
<abbr>U.S.A.</abbr>

15.4. When an abbreviation  and  a  name  coinside  then  the 
abbreviation markers are placed closer to the marked word:
<propname type=institution><abbr>USA</abbr></propname>

15.5.Foreign  abbreviations  are  coded  similarly  to   the 
Estonian ones.

15.6. Shortened forms are not marked as abbreviations:
Mul' (pro mulle)

15.7. 13-nes, 12th, etc are. not marked as abbreviations.

15.8. Chemical formulae are marked similarly  as  the  other 
abbreviations by means of the symbol <abbr>.

15.9. An abbreviation marker occurring in the  middle  of  a 
word is marked as other abbreviations:
8<abbr expan="shilling">s.</abbr> = 8 shillings

15.10. If a sentence ends in an abbreviation and in one full 
stop which may be both a sentence point or  an  abbreviation 
point then the  full  stop  will  be  included  between  the 
abbreviation markers.

16 Hyphen and dash

16.1. The hyphen (-) is used within a word  and  is  neither 
preceded nor followed by a space.

16.2. The dash is preceded and followed by a space.

16.3. A line-end hyphen  is  not  coded,  except  where  the 
hyphen is part of the normal spelling of the word (a  hyphen 
in the middle of  a  compund  word,  etc.).  Where  spelling 
practice  varies  with  regard  to  hyphenation,  a   coding 
decision has to be made as to whether the line-end hyphen is 
preserved or not. The hyphen is preserved:
a) if dictionaries show that hyphenation is normal
b) if the word in question is hyphenated  elsewhere  in  the 
same text.

16.4. - meaning "to" (as in 1632-1690) is coded  as  hyphen, 
ie not separated by spaces.

17 Mathematical expressions

17.1. Mathematical expressions are where possible  coded  as 
themselves, eg + = %.

17.2. - between numerical expressions represents "minus"
      x in numerical expressions represents  "multiplication 
        sign"
      / in numerical expressions  represents  a  divisor  in 
        fractions.

17.3. Numbers are presented without spaces between digits.

17.4. Numbers are coded by means of the symbol <num>.

17.5. The marker <num> may have an attribute  "type"  having 
the  values   cardinal,   ordinal,   fraction,   percentage, 
certainty,  and  also  an  attribute  "value"   giving   the 
numerical value of a mathematical  expression.  In  case  of 
certainty the attribute "value" is not used:
<num type=cardinal value="156">sada viiskummend kuus</num>
<num type=ordinal value="37">kolmekumne seitsmendal</num>
<num type=fraction value="0,5">1/2</num>
<num certainty>kumned tuhanded</num>

17.6. Dates are always presented without spaces between  the 
digits:
8.-12.mail

17.7. Dates are marked by the symbols <date> and </date>

17.8.   The   tag   <date>   has   always    an    attribute 
"value=yyyy-mm-dd".
<date value="1990-02-21">21. veebruar 1990</date>
Partial dates can be expressed in the "value"  attribute  by 
omitting a part of the value supplied:
See toimus <date value="08-03">3. augustil</date>
Jaan suri <date value="1940-12">1940.a. detsembris</date>

17.9. In case of date range the attribute "value"  may  have 
fillers "from to", only "from" or only "to":
<s>Soda kestis <date  value=from"1942-03"  to"1942-10">1942. 
martsist oktoobrini.</date></s>

17.10. When dates are not fixed the attribute "value"  takes 
on the filler "certainty":
<date certainty>kolmekumnendate aastate lopus</date>

17.11. Time expressions are denoted by  the  markers  <time> 
and </time>.

17.12. The tag <time> may have an attribute "value" followed 
by the time expressions in numbers.  The  following  symbols 
are used between the numbers:
hour : minute . second , fractions of a second, etc.
<time value="16:00">kell 16</time>

17.13. The results of  sports  competitions  are  tagged  by 
<score> and </score>:
<score value="4:0">4:0</score>

17.14. More complex  mathematical  expressions,  tables  and 
figures are represented by <formula>, <table>  and  <figure> 
when these expressions, tables and figures are  retained  in 
the text. When these elements are removed from the text they 
are  replaced  by  tags   <ext.formula>,   <ext.table>   and 
<ext.figure>.

18. Typographical errors

18.1. Obvious typographical errors are corrected, eg:
     Form in the source               Corrected to
         thetre                         theatre
         ofr                            for
         on eof                         one of
         for                            for
All corrections are listed under the text in question in the 
List of Text Extracts.

18.2. The following types or errors are normally corrected:
1) Errors of spelling
2) Cases of incorrect spacing
3) Duplication of words.

18.3. If there is any doubt about the intended form the text 
is left unchanged  and  the  tags  <sic!>  and  </sic!>  are 
inserted in the text. This applies to all cases of abberrant 
syntax.


18.4. The tags ,<sic!> and </sic!> are used only when it  is 
considered  necessary  for  the  user  of  the  corpus.  All 
occurrences of  this  tag  are  listed  under  the  text  in 
question in the List of Text Extracts.

18.5. Errors in the presentation  of  foreign  names  or  in 
foreign-language passage have  normally  been  left  without 
comment.