Go to:  | Biblioteket | Digitale tekster | Vestfold College | Vestfold College Library |
Tønsberg: Høgskolen i Vestfold, 2001
Egil Børre Johnsen: Textbooks in the Kaleidoscope.
| Front pageContents | < Previous chapter | Next chapter >
 
 
 
 

Chapter VII
 

Appendix:

Registration as a Basis for Analysis
 
 
 

Problems
 
 

The number of textbook titles to be examined need not be
especially large before the registration process begins to
present problems. The term "textbook" is in itself complex
and fluid. In countries with approval schemes, for example,
one has to decide whether the register is to include non-
approved textbooks. Norway's approval system has never
covered study aids and workbooks unless they are parts of
complete works, yet many of these books have played an
important role in teaching. Looking back on the past,
although many textbooks have probably been reviewed and
approved, some of them may no longer be available today. A
given title may be available in one edition but be out of print
in an extensively revised and perhaps more widely used
edition. The problem edition/impression/author/title
manifests itself in several ways when one sets about
registering or classifying textbooks from an historical
perspective. Many textbooks have been issued in several
editions and impressions; moreover, they may have changed
title, contents and author profile in the process. According to
Beckson and Ganz (1990), the term edition refers to the total
number of copies of a work printed from a single set of type
(new editions implying that changes or revisions have been
made in the original typesetting), and the term impression
refers to the total number of copies printed at one time while
the type or plates are in the press (new impressions implying
that the same type or plates are used for a reprinting). But
this distinction is not consistently observed in all countries. In
some cases it can also be difficult to obtain complete
information about the number of editions/impressions. Some
may be out of print, and practice may vary between
numbering successive editions on one hand and merely giving
the total number of copies printed on the other.
  Eva Bjørkvold and Frøydis Hertzberg (Bjørkvold -
Hertzberg 1976), who studied Norwegian grammar textbooks
published between 1816 and 1973, narrowed the field partly
by choosing representative samples of the grammars
published in many editions/impressions, i.e., grammars which
had a long life span, and partly by using formal criteria - for
example a change of title or author where information about
editions/impressions was lacking or unclear. In this way they
collocated a pool of primary source material consisting of 133
different grammars, while their total sample comprised 188
 different editions/impressions. Their presentation was based
on a purely practical solution. Regardless of the designation
on the colophon page or elsewhere in the books, they use the
term edition consistently, unless special considerations call for
more precision (p. 14).
  When the term edition is used to designate a genre, as in
school editions (of, for example, Ludvig Holberg's comedies),
there are even more sets of editions to keep track of. Which
edition of the work is the school edition based on? Has the
school edition been subject to extensive revision and, if so,
has this been indicated with a new edition designation on the
colophon page? Leif Longum has written an article on his
examination of 86 school editions of Norwegian authors
(Longum 1987) without touching on this last question. He
uses edition consistently about an individual selection's title
and impression consistently for all new reissues, i.e., a best
seller is "a work printed in 10 impressions or more" (p. 67).
Like Bjørkvold and Hertzberg, he includes the years for the
first and last numbered impressions.
  There is an important distinction between registering the
publications of textbooks on the one hand, and trying to
register their distribution on the other. The latter area is the
most difficult to survey. Even if all the publishers made their
sales figures available, it would only be possible to draw
approximate conclusions about the life cycle of books.
Second-hand books and class sets are two key words in this
context.
  For most studies, a thorough clarification of the
relationship between objectives and methods on one hand,
and the selection process on the other will be more important
than the degree of comprehensiveness. This may be illustrated
by the three criteria that Tyler Kepner (Kepner 1935, see
page 44) used as the basis for his study of the development of
geography and history books:

  (...) no effort will be made to list all titles. Rather an
  attempt will be made to select (1) those textbooks, which
  by reason of the large number of editions in which they
  were printed, were obviously influential books, (2) those in
  United States history, for example, in which the author had
  also established himself as a writer in other historical
  fields, and (3) those which merit attention for their special
  characteristics, though probably not widely used. (P. 144.)

The first criterion dominates major historical studies, while
both the last two criteria, and the combination of all three,
are unusual in such studies.
 
 

Sources
 
 
 

The usual sources are national archives, university libraries,
educational libraries and national bibliographies. Experience
has shown, however, that even all these taken together cannot
be assumed to offer complete information in the areas
mentioned above, at any rate not in the Nordic countries.
This is first of all because the system of classification will
vary over a long period of time and, second, because entries
on file cards are not always systematically and/or completely
updated. Information may also be provided by publishing
houses, insofar as they are still in existence. The largest
Norwegian publishers have some data on file about sales and
printings. The practice varies from in-depth systematic
listings in anniversary catalogs to rather more scanty records.
A common tendency is for such records to be more carefully
and systematically maintained the further back we go in time.
For example, several major publishers find it more difficult
to locate information from the period 1960-80 than from
1920-60.
  The same tendency may be seen in the filing practices of
the educational councils, relevant in this context because they
have administered approval schemes and/or made various
general and specific statements about textbooks in a number
of countries. Norway is a good example because the tradition
dates back to July 24, 1885, when the very first provisions
concerning the authorization of textbooks were issued by the
Ministry. They were followed the next year by a circular
making authorization contingent upon books following the
orthographic rules adopted in 1885. In the 1890s, several
directives were issued requiring upper secondary school
principals to list the titles of the textbooks used by their
schools in the schools' annual reports. This practice was
followed until 1916. In the first decade of the 1900s, the
Ministry was very conscientious about distributing
evaluations of new textbooks for primary and secondary
schools, based on the books' contents and educational merits.
As the number of titles increased and as the approval scheme,
which officially dates back to the Compulsory Education Act
of 1889, was further developed, the Ministry and the
educational councils were reduced merely to publishing lists
of titles in bulletins and annual reports. From 1955, such
reviews for primary and secondary schools were made public
and published sporadically in a gazette called "The
Norwegian School", which existed for 30 years. The Ministry
and the publishing houses tried several catalog schemes in an
 effort to provide up-to-date, yearly printed information
about all available approved textbook literature for primary
and secondary schools, but it became increasingly difficult to
keep track of the flood of books produced in the 1970s and
1980s. One example is the 1992 Catalog of Teaching Materials
for the Upper Secondary School (Læremiddelkatalogen for
den videregående skole 1992). It is 430 pages long and
includes roughly 5000 titles. A reference list of entries
arranged by subject runs to more than 600 words; a summary
of the various courses of study in this type of school fills 22
pages. At present (February, 1993), the Norwegian
Publishers' Association and the publishers' central
distribution company, Forlagssentralen A/S, are trying to
incorporate all titles used in primary and secondary schools in
a new computerized database system.
  In Sweden, Swedish textbooks were registered in the
Swedish National Bibliography until 1976. Since then, new
titles have been entered in LIBRIS, a communal database for
the country's research libraries. In Denmark, the National
Center for Educational Materials published comprehensive
catalogs as well as some smaller ones covering individual
subjects every year in the 1980s. However, the National
Center was closed down in 1990.
  One major source of information about textbooks in use is
the schools themselves. Bound annual reports for the two
decades just before and after the turn of the century exist for
a score of secondary schools in Norway. They are housed at
the Norwegian Educational Library in Oslo and they contain
meticulous lists of the material covered in every subject. As
previously noted, the tradition of annual reports was kept up
until 1916. In the inter-war years, most upper secondary
schools had teachers keep records of which readers and
textbooks they used and what they covered during the year.
Few of these records have survived; a survey conducted in
1990 on the subject of Norwegian elicited positive responses
and relatively complete lists from only one third of the upper
secondary schools polled in Norway (Johnsen 1992a). The
system more or less disintegrated during the 1970s and 80s,
when it was replaced by prescribed reading lists for final year
pupils. There is no rule saying that these lists have to be kept
on file. The tendency to forget is underlined here because it
should be seen in connection with a question that is discussed
elsewhere: At any given time, how much attention do
subjects/textbooks themselves pay to what has gone on before
(Kuhn 1962, see page 338)?
  A register which goes back approximately a hundred years
from 1990, and which is based on all the material available
from all of the sources named above, will on the one hand
hardly risk overlooking textbooks that have been used, even
to a very limited extent. On the other hand, such an approach
is no guarantee that absolutely every title, and certainly not
every edition or impression, will be included on the list.
  The sole objective of an investigation may be to achieve
 the most complete and detailed survey possible. Such an
objective need not only be based on the satisfaction derived
from compiling a definitive compendium. Another argument
for such thoroughness is that the existence of such material
can provide an incentive for new research. The most striking
example of this is the project "Emmanuelle" at the Institut
National de Recherche Pédagogique in Paris (Choppin -
Decouche-Beauchais 1988; Choppin - Rodriguez 1989). Every
textbook published after 1789 for every level and subject in
primary and secondary school is registered in a database, with
information about titles, authors, publishers, educational level
and edition/impression. In addition, all the information in the
database is indexed under 43 different key words or fields
("champs"), for example, illustrator/printer/library
(information about where the book is
available)/discipline/approval. There are two ways of
accessing the data; either on line access (as for all multi-
criteria data bases): 43 variable names or fields can be
combined thanks to boolean operators, and - or - not; or
through the publication of comprehensive catalogs, discipline
by discipline.
  The project is based on all the kinds of sources mentioned
above. It is very useful to have access to "Bibliographie de la
France", a journal published weekly since 1811, which lists
all new editions of books - including textbooks. But even that
is incomplete and deficient. The head of the project, Alain
Choppin, gives an account of all the sources for French
school book research in an article in Thérèse Charmasson's
basic text for and about school history research (Charmasson
1986).
  In 1991 the Emmanuelle material comprised over 80,000
titles. The results up to that time in four subjects (Latin,
Greek, Italian and German) were published in book form.
The study, started in 1979, will run for several more years.
The software is now developed to the point where it can be
adapted for use with other languages. (See page 394.) The
justification for financing such a large apparatus was and is
that the material has significance for new research. The
primary goal is to achieve the greatest possible degree of
comprehensiveness.
  When we look at studies dealing with a large number of
titles, we see that there are several avenues of approach, not
all of which place the same emphasis on accuracy or
completeness. It would be natural to assume that approaches
vary depending on the objectives of the various studies.
However, it appears that such immediately obvious
correlations do not always exist. The following survey
presents analyses based on large amounts of data that have
been organized according to the attention paid to the selection
and registration of titles; that is, the more an analysis focuses
on this problem, the farther down on the list it will come, as
a rule. The content and form of the analyses will only be
 discussed and evaluated here to the extent that they shed light
on the correlation between registration and objective. In the
main the presentation will observe the order Norwegian -
Nordic - other countries.
 
 
 

Comprehensiveness and Scholarship
 
 
 

There is not necessarily any correlation between the level of
formal scholarship and the level of justification for the
selection and registration of titles. Here I have used a doctoral
dissertation, a study done by an institute and a master's thesis
to illustrate this point.
  Karin Tarschys' doctoral dissertation on mother tongue
instruction in Sweden's secondary schools at the end of the
19th and beginning of the 20th century (Tarschys 1955) is
based on a large corpus of source material, a fraction of
which consists of textbooks and anthologies. The primary
source material consists of official documents and scholarly
and popular literature about the development of the school
system. It is against this background that she looks at the
teaching of Swedish as a mother tongue. From the preface it
is evident that Tarschys has not seen it as her main task to
deduce literary canon from her material, although that might
have been possible within the framework of such a
comprehensive survey. Her work is first and foremost an
investigation of the literature read by teachers and
researchers, not that read by pupils. It would be unreasonable
to demand a textbook bibliography under these
circumstances. On the other hand, Tarschys does not directly
address the possibilities that textbooks offer for saying
something about one of her main themes, namely the attitude
taken toward the school and society during different periods
of time. Other studies of a similar nature may reflect very
different viewpoints. In A History of English Language
Teaching (Howatt 1984), A.P.R. Howatt uses textbooks as his
primary source material and goes so far as to include short
biographies of several of the authors.
  Bjarne Bjørndal's study of textbook didactics (Bjørndal
1982) is a less comprehensive study with a lower level of
scholarly ambition than Tarschys' investigation. Bjørndal's
investigation focuses on a wide selection of textbooks. The
author has examined development trends in Norwegian
textbook literature "from about 1950 to 1981." The primary
aim of his survey is to "highlight the interaction between
newer didactic ideas, such as those expressed during the
international educational debate of the 1960s and 1970s, and
the trends currently developing in textbook literature in
Norway and certain other countries." (Preface p. 1.) Bjørndal
examines most of the theoretical subjects taught in primary
school and includes books from all grade levels. The
textbooks are indexed alphabetically by author at the end of
 the report, comprising 169 titles in all (pp. 195-209). In some
cases the number of the impression is given after the title, in
others, the number of the edition, and in still other cases,
only the year of publication. Sometimes several editions are
listed. There is no indication as to what system, if any, was
employed to order this information, nor is there any
discussion of the classification or selection criteria employed.
Nothing is said about the degree to which the author regards
the survey as comprehensive. The books are evaluated
collectively, and each is anonymous within its particular
subject area.
  The author himself calls attention to the limitations
inherent in this approach:

  We feel the need to guard ourselves by emphasizing that
  the conclusions set forth will of necessity be the result of
  our subjective evaluation. It is thus not unlikely that others
  would evaluate the registered phenomena differently. This
  does not necessarily preclude this kind of evaluation from
  having some significance for textbook writers, publishers,
  teachers and others who are required to form an opinion
  about the didactic element in textbooks. (P. 99.)

These explicit reservations about subjectivity and limited
usefulness make it unreasonable to demand a more thorough
bibliography. The work does, however, raise a few questions
regarding the problem of registration. For example, if there
are major differences between the books on a particular
subject in the area under investigation, isn't it important to
know something about their distribution, i.e., how many,
when and where? What if one textbook/one system has
dominated the market, as was the case in general knowledge
subjects and to some degree in Norwegian as well, during the
period under investigation? Wouldn't such information be
highly relevant to the stated goal of helping textbook writers,
publishers and teachers?
  Øystein Eek has written a master's thesis on what has been
written about literary history in Norwegian upper secondary
textbooks and encyclopedias during the period from 1840 to
1950 (Eek 1982). The material covers approximately fifty
literary histories and a list of references to a dozen different
encyclopedias. The literary histories, which received far more
extensive treatment in Eek's presentation than the
encyclopedia texts, have been published in many different
editions and impressions. To the extent that information
about this was available in the Norwegian National
Bibliography, from the Norwegian Education Library, or
from the publishers, Eek has included it in his bibliography.
He has thereby provided a reliable record of the pattern of
distribution within the time period and a relatively reliable
reference as regards the extent of distribution (sales figures
are not provided). This is important since Eek's study of
 educational and popular literary history writing establishes
historical viewpoints. In addition, the thoroughness of the
survey will save time and effort for others who might wish to
work with the subject.
 
 

Essay Writing
 
 
 

The fact that an investigation does not explain the
registration criteria employed or does not try to be exhaustive
is not a problem per se. The problem arises when the author
tries to draw general conclusions.
  Historian Marc Ferro (Ferro 1986) has written a book
about how history is handed down from adults to children all
over the world ("à travers le monde entier"; subtitle). The
source material consists of films, fiction, and popular and
scholarly literature about history teaching, as well as
textbooks from four continents. Ferro's presentation has no
academic pretensions. What he writes in the preface,
however, is quite fitting for the relatively modest, but
nonetheless important, tradition of larger studies of the same
type from other authors in other countries. It is important
because it deals with books that have attracted public
attention and had an impact:

  The project is so gigantic, so megalomanic, that I must
  acknowledge and justify all inadequacy.
  (...)
  Still, I have not given up the race. On the contrary, what I
  have given up is the thought of making every chapter into
  a little dissertation. A whole lifetime would not be
  adequate for that; it's as hopeless a problem as trying to
  square a circle. Before I could get to the end, I would have
  to begin again, with a new generation of books, films and
  perhaps new, as yet unknown products. (Pp. 11-12.)

Ferro gives the year and place of publication for the
textbooks he has used, but he explains neither the criteria for
nor the system of selection. The presentation remains true to
its main thought, which is to show the complex background
for how nations acquire their different images of one another.
  Another French author, linguist Dominique Maingueneau,
does the same thing in his analysis of French textbooks
during the Third Republic (Maingueneau 1979). He reviews
a classic reader (Bruno 1877) and a large number of textbooks
in language, history, philosophy and geography in order to
show how the textbooks individually and collectively describe
not only the society and the time in which they were written,
but also the school they were written for ("...le discours que
tiennent les manuels, l'école sur le "monde' (dans tous les
sens) constitue en profondeur un discours sur l'école elle-
même." (P. XV.)). The account is organized thematically in
 chapters like "L'Algérie", "Le peuple français", "Le
désert", "La mère patrie" and "Les bienfaits de la
civilisation". As with Ferro, there is no outline for or
systematization of the selections. The titles are presented
exclusively, and quite randomly, in the form of footnotes.
  The approach has been characterized as "impressionistic"
(Togeby 1978), based on another work which also follows a
distinct train of thought: Herbert Tingsten's study of ideology
in textbooks, particularly those for history, geography and
religion, published in a number of countries (France, Italy,
Germany, the USA, Austria and Sweden) during the past
century (Tingsten 1969). But Tingsten provides no systematic
overview either. The dust jacket of the Norwegian edition
describes the work as "a very comprehensive, dedicated and
scholarly study." But that description does not hold up when
measured against the standards of academic reliability and
validity that would commonly be required in connection with
content analyses of the type Tingsten has engaged in here.
  Although Frances FitzGerald (FitzGerald 1979) shares
Tingsten's ideological starting point, she is also very
concerned with demonstrating how commercial interests play
a part in determining the content and viewpoints of
textbooks. Yet she does not explain how textbooks included
in the study are selected either. Her only form of
systematization is a bibliography where textbooks are listed
together with other titles under the heading "United States
History Texts", chronologically arranged in five different
time periods.
  Besides "impressionistic", the writings of authors like
Ferro, Maingueneau, Tingsten and FitzGerald can also be
characterized as essays. They are included in a survey of
textbook studies because, first, they deal with a large corpus
of material, second, they represent central approaches to the
study of textbooks and, third, these titles, with the exception
of Maingueneau, have become best sellers which have had a
profound impact on the public's perception of the nature and
significance of textbooks.
  Several surveys of textbook research have tended to omit
inquiries written con amore, accounts which primarily aim at
describing educational developments, content and culture-
building endeavors in textbook literature. One early example
is Clifton Johnson's richly illustrated journey through
American textbook literature of the 18th and 19th centuries
(Johnson 1904; new edition 1963). His point of departure for
writing the book was very nostalgic at the turn of the century:

  The contrast between the dainty picture books that are
  provided to entice the school children of the present along
  the paths of knowledge, and the sparsely illustrated
  volumes conned by the little folk of two or three
  generations ago, is very great; and yet the old books
  seemed beautiful to the children then, and the charm all
  comes back when a person of middle age or beyond
  happens on one of these humble friends of his youth. What
  an aroma of the far-gone days of childhood hovers in the
   yellow pages! (P. XVII.)

Johnson handles the material with the same freedom as, for
example, Ferro and Tingsten, but he provides rather more
detailed information about which institutions in which states,
which individuals, which book dealers and so forth he has
consulted. Among his sources is The Henry Barnard
Collection, which boasts a virtually complete collection of the
textbook titles used in the USA prior to 1850.
  Studies based on one particular book collection are rare.
John A. Nietz' Old Textbooks, which looks at American
school books from colonial times to the year 1900, is a good
example, however (Nietz 1961). The author himself collected
the books, which comprised over 8,000 volumes when he
began to write about them. He was motivated by the thought
that "An analysis of the school textbooks used in the past
reveals a truer history of what was taught in the earliest
schools than does a study of past educational theories alone."
(P. 1.)
  Carpenter's American textbook history (Carpenter 1963)
begins in the 1700s and continues up to modern times, but its
main emphasis is on the 19th century. Like Marc Ferro and
Karin Tarschys, Carpenter has compiled a bibliography
dominated not by textbook titles, but by books about school
and cultural development. He is as free as Johnson in his
presentation, but whereas Johnson delights in highlighting
pages of text and illustrations, Carpenter is more concerned
with textbook writers and the history of publishing houses.
He also stresses the development of the subject American
language and literature, although without specifying any
survey of titles or direct selection criteria: "No attempt has
been made to present a comprehensive bibliographical
treatment of the subject, (...) It is hoped that no outstandingly
important texts or classes of texts have been inadvertently
omitted." (Pp. 7"8.)
 
 

Historical Investigations
 
 

Readers
The standard work in historical textbook research in the USA
is Ruth Miller Elson's study of American textbooks in the
nineteenth century (Elson 1964). She scrutinized more than
1,000 of "the most popular textbooks used in the first eight
years of schooling, since high school was not then a normal
part of the education of most Americans (...)" (p. VIII). This
work distinguishes itself with an unusual thoroughness in the
references, except that it contains no information about how
the titles were obtained. They are arranged chronologically in
five subject groups: primers, readers, and geography, history
and arithmetic books. Only one date is given for each title
(the year of the first edition), and the authors' and publishers'
names are included. Every title has a code number, and the
author uses the codes, with page references to each individual
book, in the form of footnotes in the body of the text. These
footnotes cover an average of one-fifth of every page. When
she writes, for example, about descriptions of Columbus, the
number of source references on a single page runs to over
thirty (p. 190). Although the situation is hypothetical, this
means that a reader with access to the books could check
every single reference.
  Readers dominate Elson's investigation and several of the
others mentioned previously, as well as the many historically
oriented surveys and analyses in German. Two of the newer
standard works are Heinz Tischer's and Hermann Helmers'
descriptions of the history of the reader in German primary
school (Tischer 1969, Helmers 1970). The former covers the
time up to the mid-1800s and, generally speaking, the latter
covers the period 1650-1967. A comparison of the two
approaches shows that registration and selection can have a
significant impact on the final evaluation and results.
Helmers' method is discussed on page 92; he is selective and
uses angles of approach based on didactics and cultural
history. He sets up different perspectives and arrives at
different conclusions from Tischer, who stops at around 1850.
This is not because there were too many titles. In contrast to
Helmers, Tischer claims that up to that time readers had
undergone constant development and renewal as agents of
social criticism. According to Tischer, the topical portion of
their educational function subsequently disappeared from
textbooks which, in spite of changes in text selection and
equipment, no longer undergo any real development, but
remain locked in the role of the conveyor of traditions.
  The decisive criterion for Tischer is originality, based on
the conflict between tradition and renewal. His primary aim
 has been to investigate what happens to the "old" in the
books which he and/or his predecessors regard as "new."
Tischer refers to three major studies as the foundation for his
own work, pointing out that they have each had their own
way of dealing with large corpora. One has tried to register
and describe everything, but singled out a few titles for more
detailed analysis (Fechner 1879); a second has also selected a
few titles, which make up the entire corpus (Krumbach 1894
and 1896); and a third used many titles that are evaluated
separately according to an extremely fragmented system of
categorization. (Bünger 1899.)
  The best known German study of the reader's history in
schools of higher education was written by Peter-Martin
Roeder (Roeder 1961). He registers more than 100 titles from
1700 up to the present day and includes a list of secondary
literature covering even more titles. In his introduction,
Roeder remarks on the lack of previous work relating to
reading books for the higher grades, asserting that
distribution has been difficult to ascertain. He claims that the
development of reading books in the 18th and 19th centuries
varied considerably in different parts of Germany, but he
does not have enough material to document the claim.
Nevertheless, Roeder shares Tischer's view that, generally
speaking, readers became more alike in the latter half of the
19th century, a factor which justified highly discerning
selection procedures.
  It is notable that Tischer's three leading references all deal
with the previous century. The bibliographies in large-scale
historical surveys by other researchers (Carpenter 1963, Elson
1964, Choppin 1977) may also suggest that comprehensive
registrations/analyses with historical aims were more common
in the second half of the last century than in our own times.
  Germany has an exhaustive production of studies, books
and articles on readers. Dieter Marenbach has put forward a
number of theories on the reason for this in an analysis of
readers from the 1970s (Marenbach 1980). Until well into our
century, readers were also the most important subject texts.
Most adults have more positive associations with this first
school book than with others. Readers can be used more
independently of the teacher, and they tend to promote
reading more than most textbooks. Of all the printed
materials used by schools, readers have traditionally been
considered the most important transmitters of values.
  The bibliography for Marenbach's article refers to 23
readers and to 51 different books and articles about the
subject. This phenomenon, i.e., that there are more references
to theoretical studies than to primary sources, is not limited
to German literature on the subject. It is also internationally
common in studies on the general history of literature, for
instance. Yet in the case of readers, where the quantity of
material in German is so enormous, the discrepancy is
striking.
  Reference is also made to Marenbach's article in this
context because it demonstrates how difficult it is to
 systematize the literature on the topic. Including subtitles, the
51 references discuss historical as well as educational,
didactic, methodological, psychological, ethnographic and
political approaches - which may overlap in individual titles
or through several titles.

Grammar Books
Apart from the Emmanuelle project, French textbook
researchers cannot point to as many large-scale surveys as
their German colleagues can. One investigation that does
indicate problems and discuss boundaries is André Chervel's
analysis of French school grammars during the past two
hundred years (Chervel 1977). He demonstrates the size of the
corpus by giving a complete listing of the grammar titles
published in the course of a single year, 1840. Most are school
grammars, and there are a total of 34 titles. Altogether, he
registered 2500 titles used in the 1800s, most of which were
designed for use in schools. He has built upon a large survey
by Jean-Claude Chevalier (Chevalier 1968), who, after
analyzing grammars up to 1750, writes that "a systematic
registration of all these titles is beyond the reach of a single
person." (Chervel, p. 29.) In a note discussing the selection,
Chervel describes his own method for sifting through the
large body of material:

  I had to be satisfied with a typical selection. This was
  actually self-evident. The material is, from book to book,
  characterized by repetitions ("caractère éminemment
  répétitif"); plagiarism appears constantly among textbook
  authors. In my historical survey I have tried to register and
  date all editions that represent changes or innovative
  thinking, but this will always be an approximation which
  more detailed research could modify. On the other hand,
  I feel that repeated observations of the same phenomena
  within given time frames will be sufficient to place the
  advent of new trends or schools of thought within a
  definite period. (P. 29.)

Further, Chervel points out the problem that all textbook
researchers run into and which appears to increase in direct
proportion to the size of the corpus: Ideally, the complex
function of textbooks and the overlapping nature of school
subjects ought to be supplemented by other views. What
about related subjects like grammars in Greek, Latin and
modern foreign languages? What about language books which
include far more than just grammar? What of the
distinction/connection between the renewal of grammar
theory on the one hand, and the renewal in the textbooks'
educational methodology on the other? How do they compare
with grammars in neighboring countries - with the same or
a different language - during the same periods?

Anthologies
In many countries the twentieth century saw more widespread
 distribution of selected texts, or anthologies, which expanded
the selection of readers available to children and young
people. It is difficult to summarize anthologies because they
have so many authors and so many kinds of contributions.
This is particularly true of the registration of modern
anthologies compiled after WW II, because so many have been
made and they are so comprehensive. Still, the Nordic
countries have made a few studies that attempt to systematize
large corpora of anthology material.
  For a variety of reasons, Torben Frische's investigation of
the reading of literature in the Danish upper secondary school
from 1910 to 1971 (Frische 1977) focuses on individual
works. The author points out that individual works dominate
syllabi in terms of the number of pages, and that the sources
(syllabus records and yearbooks) give more reliable
information about principal works than about individual texts
contained in or separate from anthologies.
  Frische's work is subtitled "The Promotion of Cultural
Literacy in Grammar Schools as Illustrated by a Statistical
Investigation of the Texts used to Teach Literature". The
scope of the information obtained, the discussion of the texts'
selection and use, and the overall tabular presentation is the
most thorough model covering the use of historical material
from the individual schools available in the Nordic countries.
Frische uses information from the syllabi of a total of 1,470
final-year classes during the period in question. This results
in a total of 21,870 pupils, who studied 73,877 principal
works included in 177 different titles.
  The existence of such a large corpus combined with his
own special angle of approach created certain problems for
Frische, something which indirectly raises the question of
what is ultimately more important, the registration or the
analysis. About this he writes:

  But now that this material, which I am using for a special
  purpose, has been sought for other purposes and, as I also
  pointed out in the introduction, will be available for
  further use, I see no particular reason to let my report be
  governed by the exclusive use I have made of it in this
  dissertation. (P. 38.)

One Nordic anthology study that is remarkably thorough is
Staffan Björck's survey and analysis of Swedish poetry
anthologies from 1730 to 1983 (Björck 1984). Björck
examines nearly 400 titles but does not include school
anthologies. However, the work covers many principle
viewpoints about the anthology as a genre, and it is
mentioned here not least because of its comprehensive list of
titles. The catalog, prepared by librarian Sylvia Törnkvist,
contains all the standard bibliographical information about
the editor, year, impression, revisions, number of pages and,
in addition, individual treatment of the authors and contents
in the form of short personal comments. In addition to this,
every title is marked with symbols that give information
about the way in which the anthology in question was
organized. Viewed together, the symbols represent a register
which not only expands the normal scope of registration, but
 goes a long way toward forming a pattern to describe
anthologies as a genre. A total of ten such categories are used,
reiterated here in the form of key words, as in the book (the
symbols are not given). The key words:

  - indicate that the anthology is limited to a particular area;
  - that the publisher has attempted to make a
    representative selection that covers the material fully;
  - that the selection is rather personal, subjective; there are
    many fuzzy areas between the extremes, but most
    anthologies are arranged either chron(ologically),
    alph(abetically) or them(atically);
  - indicate that elementary data on the authors has been
    provided;
  - indicate that detailed information about the authors has
    been provided;
  - indicate that poem sources have been provided;
  - indicate that other bibliographical information has been
    provided;
  - indicate an artistic lay-out made solely for the work in
    question;
  - indicate photographic illustrations;
  - indicate a portrait of the author irrespective of the
    technical execution.

Mother Tongue Language and Literature Books
Ian Michael's book, The Teaching of English (Michael 1987)
is a standard work both in this group of books and in general.
Textbooks in English language and literature comprise most
of the material used in a study of mother tongue instruction
throughout three centuries, from approximately 1570 to 1870.
Michael had a dual purpose. He wanted to refute the assertion
that the subject of English was "a relatively recent addition
to the curriculum" (p. 1), and he wanted to show what a
significant corpus textbooks in particular comprise as far
back as the 16th century. The result is a bibliography of 2,708
titles (of which 183 are American); of these about 70 per cent
were consulted and used directly in the analysis.
  Nevertheless, in his introduction Michael writes that his
registration and analysis are only a beginning; he is
undertaking a "tentative quantitative analysis" of the
material and has had to use "marginal judgments about what
is relevant" (p. 6). Michael's definition of textbooks is very
broad. It covers books used by pupils at school, books which
are read aloud in class as part of schoolwork, books used by
teachers and parents for guidance and handbooks/instruction
manuals. Categorization represents a major problem. The
author has managed to limit the categories to a total of ten,
but he emphasizes that the category transitions are often
fluid. Categorization is more difficult in literary than in
language disciplines. Each category has a code which is listed
after the year in the bibliography. The ten categories are
listed here to illustrate the problem:
 1   S   Spelling, excluding orthographical reform;
           dictation; elementary punctuation; exercises in false
           spelling.
2   RE  Elementary reading.
     RA  Anthologies; single-author texts; readers not in
            series; copybooks.
      RH  Books primarily for home reading but perhaps used
             in school.
     RS   Readers in series.
3   P     Performance; elocutionary and dramatic texts;
            pronunciation; memory.
4   B    Belles lettres; rhetoric; criticism; prosody; history of
            literature.
5   G    Grammar; exercises in false syntax; advanced
            punctuation.
6   La   Language; vocabulary; orthographical reform;
            etymology; history of the English language.
     D    Dictionaries used in school.
7   Ex  Written expression; oral expression when
           distinguished from performance; debate.
8   C   Compendia.
9   Ed  Educational texts relating to English teaching.
10  Lo  Logic: texts likely to have been used in school.
(Michael 1987, p. 9.)

Like Bjørkvold and Hertzberg (see page 354), Michael does
not distinguish between edition and impression, but chooses
the same solution from practical necessity: "The term`edition'
is used here for all printings and is not distinguished
from`impression'." (P. 386.) Sources are given for the books
that have been located. Not found and not analyzed books are
included in the survey if they are "potentially significant
works" (p. 386). The number of pages in a book is noted only
if it is conspicuous for being particularly high or low. The
publisher is only listed when the information serves to
distinguish between similar-sounding titles.
  In the article "Aspects of Textbook Research" (Michael
1990), the author has summarized his experiences, principles
and desires in this way:

  Some implications for research

  (i) The decision whether or not to treat a publication as a
  textbook will have to be made for each work separately.
  Criteria will include: author's explicit or implied intentions;
  publisher's advertisements; publishing history; use made of
  the book; price; tone of the book (instructional,
  explanatory, moralistic).

  (ii) The publishing history of a book provides the best
  factual basis for an assessment of its influence.

  (iii) We need bibliographical studies of important
  textbooks, including the identification of editions which
  contained significant changes.

   (iv) We need to collect available and seek out fresh,
  information about the location of copies. Textbooks are not
  often cataloged by the libraries which hold them.

  (v) It would be useful to collect biographical,
  autobiographical and fictional references to particular
  textbooks (Cobbett as a young soldier learning Lowth's
  grammar by heart).

  Only when we have a far wider range of factual
  information will it be possible to bring the evidence of
  textbooks to bear on aspects of social and political history
  which they could illuminate. (Michael 1990, p. 7.)

History and Geography Books
Textbooks make up the main corpus of material for the two
largest Nordic studies of the development of history teaching.
Göran Andolf has written a dissertation on the period from
1820 to 1965 in the Swedish upper secondary school (Andolf
1972), and Jørgen Møller has written about Danish upper
secondary school teaching over the past 100 years, with
emphasis on the years after 1930 (Møller 1983).
  Göran Andolf's survey of Swedish textbooks through 150
years is based on the most thorough registration of history
books carried out in Nordic textbook research up to the
present time. In a separate chapter entitled "The Textbook
Stock", Andolf gives an account of the limitations and
possibilities inherent in the work of compiling a survey. A
comprehensive filing system is not necessarily a guarantee for
exhaustiveness in the sense of encompassing all published
titles. Already in 1817 Lorenzo Hammarsköld published a
"directory" of all the textbooks published in Sweden up to
that time. Hammarsköld's work was based on the yearly
summaries from book printers, but some of these summaries
proved to be incomplete. Beginning in 1817, these efforts
were systematized by the so-called "textbook audits" which,
up to 1843, listed 77 textbooks in Swedish and 107 in history.
It has since come to light that nine titles were missing from
that survey. In contrast, the registration which was continued
until 1868 by the "textbook commission" is believed to be
entirely reliable. From 1877 to 1910, textbook surveys were
carried out centrally as a part of the work of Sweden's
Official Statistics. This registration was very accurate for the
first three years; then it was decided to make surveys every
tenth year rather than annually. It was deemed unlikely that
a new title would come on the market and then completely
disappear from it in just ten short years.
  The question of comprehensiveness is also relevant to the
type and amount of information given about each book.
Where such key questions as grade level and distribution are
concerned, the sources are inadequate. The two sources
mentioned above (the audits and the commissions) include
books with subtitles such as "For the School and the Home"
and "For Ladies and Young People". It is not known whether
these were regarded and/or used as textbooks. The audits
refer to yearly reports from school principals, but repeatedly
point out that such reports were often very summary.
   In 1856 it was decided that all schools should issue annual
reports including information about the textbooks they used.
Andolf writes that from 1859 to 1961 yearly reports were
printed from "nearly all upper secondary schools" (p. 118).
The textbooks are as a rule listed there, although often only
by title and possibly one author's surname.
  The National Textbook Committee was established in 1938
and, as a result, lists of all approved textbooks were made
from 1940 to 1955. Since then booklists have been published
regularly in "News from the National Board of Education".
  Andolf concludes the survey of his possibilities with the
following summation:

  We have no complete list of all the books used in upper
  secondary schools from 1821 to 1843, but we know more
  about the period from 1876 to 1910. We know which books
  the schools have had to choose from since 1940, but not
  which were actually chosen. This is of little consequence
  for the general history book used in upper secondary
  schools, since there was usually just one textbook and its
  successors to choose from until 1954. It is hard to say
  whether, and if so, to what extent, non-approved textbooks
  were used in the classroom. It is unlikely that an
  examination of the annual reports would help clarify the
  situation. (Pp. 119-120.)

On the basis of the source material up to 1910, Andolf has set
up a table showing how frequently the books were used. He
too uses spot checks for every decade, with a few minor
exceptions and one major one: From 1930 to 1957 the
selection was so limited and the changes so small that a closer
examination was superfluous. The table shows how many
upper secondary schools used the textbooks that were
available in 1844, 1860, 1870, and so on up to 1961. The table
gives a conclusive basis for determining which titles
dominated during the period.
  It is not clear, however, whether individual schools used
more than one title, something which according to Andolf
himself (p. 129) is the rule after about 1940. Nor does the
table say anything about which editions were used; the author
consistently refers to the year of the first printing.
  As regards book size, Andolf uses a system for converting
to standard pages (p. 131). The system involves converting the
text of each book into 2000-character pages, and it offers a
highly reliable way of comparing books. The fact that Andolf
has counted illustrations as ordinary text presents more of a
problem.
  There are certain limitations in Andolf's selection.
Teaching and textbooks in national (Swedish) history are not
included in the study. And "Books which are never or only
rarely used in upper secondary schools are not discussed, nor
are readers." (P. 117.)
  Jørgen Møller (Møller 1983) deals with history instruction
in the upper secondary school from 1900 to 1972. His point
of departure is partly the desire to investigate the view of
society and the world presented in the classroom and partly
to look more closely at the relationship between history as a
 scholarly discipline and as a school subject. His presentation
is based on three groups of sources: public documents
("formalia"), textbooks, and historical source material.
Counting a 55-page tabular survey of the distribution of the
books, the discussion of textbooks comprises 150 of the
book's 261 pages.
  Altogether, the textbooks cover 26 versions of world
history and 6 versions of Nordic history. The books are
presented consecutively one by one, and accurate information
is provided about editions, years of publication, and changes
in authorship. The schools' programs and annual reports are
listed as sources. Møller has registered distribution and use
throughout the period, based on steadily decreasing intervals
as changes began to occur more quickly and the number of
titles increased (1900/1920/1930/1940/
1950/1955/1960/1963/1966/1969/1972). In 1900, 31 schools
were investigated, in 1972, the number was 106. The selection
of schools aimed at achieving a cross-section of sizes, types
and locations. In this way the author has produced a picture
of the distribution and life span of every title throughout the
entire period.
  Another study of history books, Staffan Selander's
Textbook Knowledge (Selander 1988), lists 42 textbook titles
as the basis for the analysis. In contrast, Selander provides no
information about the books but the name of the author, the
grade level and the year of first publication. His point of
departure and his methods are totally different from Andolf's
and Møller's. Selander wants to investigate "how knowledge
has been structured and how it can be understood and
explained through a pedagogical text analysis." (P. 9.) He
selects two topics - the French revolution and Sweden's years
as a major power. He then looks closely at the presentation of
the topics, mainly in two books, but with examples from and
references to the rest of his material. He offers no initial
detailed explanation as regards the temporal or geographical
distribution of the books. On the one hand, Selander can refer
to past research in this area, and one might also say that his
reasoning becomes apparent as his discussion progresses. On
the other hand, one might ask whether more detailed
groundwork would actually have changed the validity or
nature of the analysis.
  Lena Olsson has studied the view of culture presented in
Swedish geography books used in lower and upper secondary
schools from 1870 to 1985 (Olsson 1986). Like Göran Andolf
and Jørgen Møller, she has compiled a large corpus, well over
50 titles if one includes editions for different grades. Like
Andolf she has discovered a great deal about distribution with
the help of the documents from the audits and commissions
from 1876 to 1910. Olsson's study includes a thorough
accounting of sources, publication years and editions (pp. 76-
83). Distribution is the operative factor in this statement from
her presentation:

  "The question of which textbooks ought to be included in
  my study cannot, however, be divorced from the
   underlying assumption that the contents of teaching
  materials also influence the teachers and pupils who use
  them. The consequence of that assumption is that the
  textbooks used more frequently in the schools ought to
  form the basis for the analysis (my emphasis)." (P. 13.)

This may be viewed in conjunction with Olsson's point of
departure, which is distinguished from Andolf's, Tarschys'
and Møller's on one side, and Selander's on the other. The
first three attempt to use their corpora to capture a whole
teaching tradition. Selander is intentionally selective because
his goal calls for a detailed analysis, partly based on language,
of what it takes to create an educational knowledge-bearing
text in the subject of history. Like Andolf and Møller, Olsson
analyzes the contents of the books. But she is selective in two
ways: Her topic is limited to the view of culture presented in
the books, and her choice of material is determined by the
distribution of the books. Her study is based on two
assumptions: that textbooks have a strong influence, and that
this influence is both quantitative and cumulative. Such
assumptions are also presented by Andolf and Møller, but
they do not accept the consequences of them to the same
degree. Torben Frische stands at the other end of this scale
(Frische 1977). In his case one might be justified in claiming
that the registration itself, including the discussion of and
attempt to achieve a completely representative cross-section,
is a major issue. But if so, it is one of several major issues. It
is important to establish that during the 1970-1971 school
year, 32 of 100 Danish upper secondary school pupils read
Erasmus Montanus. But it is also important to know that one
of 100 read Peer Gynt. Minimal representation, or absence,
may say something in itself. In addition we learn, for
example, that each of these two works are grouped under two
different types of literature; "critical neo-humanism" and
"dialectical idealization". Compared with other empirical
material, this information achieves perspective.
  Raymond Humbert's commemorative book about the
French revolution (Humbert 1989) offers a combination of
documentation and analysis, although it is more
impressionistic in its avenue of approach. The author sees the
revolution as it must have appeared to French pupils from
books published between 1881 and 1936. Since he places a
great deal of emphasis on illustrations, availability and
technical quality played an important part in Humbert's
choice of titles.
  Alongside the strong tradition of reader research, there is
another tradition that has been prominent in German
textbook research. Its center is still located at its point of
origin, in Braunschweig, where the Georg Eckert Institute for
International Textbook Research was founded in 1951. The
objective was to advance peace and understanding among
nations by examining history and geography textbooks in
various countries, with particular emphasis on how each
presents its own country and other nations. Naturally,
 national groups selected the books from their respective
countries, and they chose widely distributed books which
were currently in use at the time. More important than an
historical survey (of the textbooks) and comprehensive
registration was a close reading of the most commonly used
texts about their own and other countries.
  Otto-Ernst Schüddekopf summarized the experience gained
from the work, which was conducted at the institute in
Braunschweig and elsewhere, primarily under the direction of
the Council of Europe (Schüddekopf 1967). Schüddekopf
does not discuss the problem of registration and selection.
However, in one place he quotes a statement from one of the
multilateral discussion meetings: "The trouble with textbooks
is that with rare exceptions they are all alike." (Eugene N.
Anderson; p. 150.) Indirectly, this may be part of the reason
why the problem of selection is avoided in so many
discussions of the subject, i.e., there may be some sort of
unofficial consensus that if you know one, you know them
all. If so, this has not been stated explicitly by those who have
actually undertaken the work of examining large quantities of
books in a given subject.
  The survey presented thus far in this chapter has given a
certain impression of the growth of early textbook research.
The field began to expand during the interim between the
world wars and experienced a strong surge in the first decade
after WWII. However, this in no way implies that textbook
registration and analyses were non-existent before well into
the 20th century. Significant, if relatively exceptional, studies
were conducted very early in the century. One well-known
example is W. Lietzmann's work on the teaching of
mathematics in upper secondary schools in Prussia from 1880
to 1906 (Lietzmann 1909). Among other things, the author
examines Kambly's Elementary Mathematics, a textbook first
published in 1850, which had sold over 850,000 copies by the
turn of the century.
 
 

Investigations of Contemporary Literature
 

There are two particularly thorough, non-historical Nordic
surveys that deal with larger groups of textbooks evaluated in
their entirety (in contrast to linguistic analyses, for example,
which might be based on text samples from an even greater
number of titles). The first is Mogens Jansen's registration
and analysis of Danish readers for the first through the
seventh school years (Jansen 1969), and the second is the
Swedish Ministry of Education's Report from the Teaching
Media Survey (Ds 1988:22-24).
  Mogens Jansen's survey covers all the books in use in
Danish primary and lower secondary schools in 1966, a matter
of more than 20 different series of textbooks, several of
which consist of more than a dozen titles for various grades
and levels.
  The registration was invariably carried out by two groups
working independently of one another. The combined result
is first and foremost a comprehensive bibliographical survey
in the traditional sense. In addition, it is a work which,
through practical examples, demonstrates problems and
solutions for those who wish to use the registration as a basis
for content and genre analyses. The object has not been to go
so far as to evaluate the subject content and methodology in
the books, but to "investigate whether, by registering the
subject matter in readers, and perhaps to some extent follow
up by a reader analysis, one might find some interesting
starting points for registering teaching activities defined on
the basis of subject matter. This did not prove possible
through these registrations, although in the process one
skimmed a number of subject areas, and worked in a rather
detailed fashion with others." (P. 12). Examples of
registration fields that were studied closely are ways of
categorizing material, ways of calculating standard page sizes
and ways of differentiating between original (what is new or
reworked for the textbook) and old material.
  In a series of annotated tables, Jansen shows how such a
comprehensive registration can be used for analyses based on
a number of different approaches, e.g., studies of authors'
canons, of the relationship between the material in readers
and that in textbooks for other subjects, of genre and of
guides and handbooks. In addition, Jansen discusses the
problem of measuring texts in historical and geographical
perspectives. In this context, the study emphasizes the ratio
of Danish to non-Danish material. The results are presented
in a figure which compares different countries using a scale
 on which the average number of pages devoted to Danish
material is valued at 1, and in which country size has been
taken into account. Not unexpectedly, Denmark fills a greater
portion of the map than Europe, Asia and Africa combined.
This figure has since turned up in textbook studies from
other countries.
  In an introductory chapter, "Collected descriptions of
textbooks" (pp. 13-16), Jansen gives a brief survey of earlier
literature on the topic. Firstly, he contends that little has been
written about it, at any rate in the Nordic countries. And
secondly, he points out that any attention devoted to the topic
at all has more frequently been occasioned by literary,
essayistic or polemical texts than by textbook analyses:

  "The transition from rather emotional debates such as
  those mentioned here to a far more measured, controlled
  debate can scarcely be accomplished without initiating
  registrations and analyses, which may serve as a -
  sometimes disturbingly pedestrian - control of otherwise
  well-formulated views." (P. 16.)

In closing, Jansen examines the possibilities for using the
study as the basis for evaluating the quality of readers. He
dismisses such a possibility on the grounds that such an
evaluation "only really has meaning when undertaken within
the framework of the individual school by the individual
teacher in one specific class and with a view to that class's
individual pupils." (P. 265.) Jansen points out that the
systematization of all the books available in one discipline
would ensure reliability in the choice of books; a choice "that
is otherwise very difficult for the individual teacher and
scarcely possible for the entire teaching staff of a school." (P.
13.) This point of view provided the primary motivation for
a more comprehensive registration and systematization
carried out by one of Jansen's co-workers: Danish
composition books for use in grades 8 - 10 (Christiansen
1985).
  In 1987 Sweden's Ministry of Education was assigned the
task of "surveying the teaching material situation in the
school" (Ds 1988:23, p. 7). The results were published the
following year in the form of a three-volume report (Ds
1988:22, Ds 1988:23, Ds 1988:24). Volume 2 (School books 2 -
 400 primary and lower secondary school classes) is an
account of a registration carried out in a nationwide selection
of municipalities, schools and classes in cooperation with
Sweden Statistics.
  The municipalities were chosen and the results processed
through the use of advanced mathematical statistical theory.
The objective was to obtain data that was so representative
that the responses from the 376 classes in 138 schools from a
total of 40 municipalities could give results that were valid
for the country as a whole in all the areas investigated. The
municipalities were divided into five types (urban centers,
 service centers, industrial centers, sparsely populated areas
and average municipalities.) In processing the statistics, the
income levels, political situations and immigrant populations
of the individual municipalities were also taken into
consideration.
  On the one hand, the point was to measure the conditions
surrounding the use of textbooks (finances, purchasing
practices, library situation) and, on the other, to look at the
way in which textbooks are used (on their own, or as
supplemented or supplementary materials). As regards
registration, the study is quite unique in its thoroughness. The
first of 40 tables gives a detailed summary of all the titles, in
every subject and at every level, in use in the 138 schools in
the course of the 1987-1988 school year. In addition, the
percentage of occurrence of every title in the various classes
has been calculated, at the same time as the study projects the
same percentage estimates for the nation as a whole.
  Because the target was contemporary, because the initiative
came from the school authorities (so that the municipalities
followed up) and because five employees at Sweden Statistics
were available to help, this registration from the 1987-88
school year represents a new development. It is also
innovative in that it registers both titles and use. Moreover,
it verifies observations made in a number of areas by studies
from other time periods based upon assumptions drawn from
circulation and sales figures alone: First of all, despite the
tremendous selection of textbooks from which to choose
today, a few titles dominate in every subject. And second,
textbooks play - at any rate measured in terms of time
consumption - a major role in the teaching of most subjects.
  The Swedish Ministry of Education commissioned a
number of other textbook studies prior to this one. The first
examined production conditions in the textbook market (Ds
U 1978:12, Ds U 1978:13) and was based on a registration that
formed a thick catalog of books currently in use at the time.
Another investigation looked at teaching material coverage in
different municipalities (Ds U 1978:14).
  Still another Swedish study was entitled The Function of
Teaching Materials in Instruction (Ds U 1980:4, Ds U 1980:5).
The objective, and thus also the registration, were entirely
different from the 1987-1988 study. The mandate was "to
study the extent to which teaching materials control
instruction and the conditions under which such control can
be avoided or reduced." (Ds U 1980:4, p 1.) An examination
was made of selected subjects and how they were taught,
based partly on observing classes and partly on interviews.
The sample of schools and municipalities was far more
modest than in the 1987-88 study, and there were so few
textbooks in some subjects that the references to them could
be contained in a single footnote. Nor was any systematic
distinction made between different types of textbooks, for
example basic texts and workbooks. The tables list textbooks
under the heading of "printed text material", in contrast to
other types of teaching aids, such as radio programs or
 stenciled hand-outs. In subjects with a larger selection of
textbooks, such as Swedish, booklists were made, including
titles only. No indication was given of when the books were
published or how widely distributed they were. This is not
unusual since all such information is available in the study
from 1978. On the other hand, it is evident that in the subject
of Swedish, for example, the individual school used different
Swedish books, representing the following types of textbooks:
readers, reading skills, workbooks, basic texts, fact books,
keys, penmanship, guides, wordlists, experience, expressions.
Reports were made on 24 of a total of 29 "observations" (1
observation = One 40-minute class). The reports concurred
with the summaries for the other subjects investigated, i.e.,
that the use of "printed text material" occupied an average
of half the teaching time. The study evaluated the control
effect of such material on the basis of five categories:
methods, material selection, language, ideology and learning
effect. One question that arises concerns the degree to which
one can draw any conclusions in these areas - about
"ideology", for example - without a registration of the use
of "printed text materials" that systematically differentiates
between, e.g., wordlists and passages on literary history in a
basic text. The study also distinguishes between two other
levels: "Information-bearing supplementary materials" and
"non-information-bearing supplementary materials". In the
first group, there is a further division between "central
teaching materials" (which include "printed text materials")
and "other supplementary materials". It may appear that the
books - which may primarily have been the main point of
departure - receive less attention as genres and isolated
variations as the study probes deeper into the manifold,
complex process that a classroom teaching hour is. The
group's research work develops almost of natural necessity
from a relatively product-oriented starting point toward a
more process-oriented position. This also brings to light a
problem discussed rarely if at all in textbook studies from the
1960s and 1970s, that is, the question of how broad to make
the definition of the concept "textbook".
  Up to 1989/90, the registration of textbooks in Germany
was a problem due to the boundary between East and West.
Many West German titles were not known or available to the
people in East Germany who might have wished to examine
them. Such examinations were undertaken in East Germany,
although on a very modest scale and undertaken mainly for
their deterrent effect. They were comparative studies of an
ideological nature, a fact that did not prevent East Germany
from also conducting extremely advanced theoretical textbook
research (Schulbuchgestaltung in der DDR 1984). In West
Germany, the registration of East German books was simple
because in the East most subjects in primary and lower
secondary school had only one textbook. The authors were
always multiple and unknown; the title pages often indicated
no more than "Autorenkollektiv". Textbooks in history for
the first six grades, Geschichte 5-10, were examined by West
Germany's Reinhard Sprenger in an article about the two
Germanies in 1977 (Sprenger 1977). In the reference to
 volume 10, Sprenger makes the following note: "Geschichte
10 - different authors, who are presented with all the titles -
Volk und Wissen, Volkseigener Verlag, Berlin 1973."
Sprenger is no less dismissive of the picture of Germany
presented in books from the East than East German analysts
are to the representation in the West. By way of introduction
(p. 82), Sprenger questions the point of analyzing a textbook
that has a monopoly in a particular subject. He refutes the
validity of such an analysis, contending that a selection of
several competing titles ("plurality") is a condition for what
he calls the conveyance of true history. In this context he
draws attention to the multiplicity of West German titles. Yet
he never questions whether having a wide selection of titles
guarantees the presentation of more than one version of
history. This has been pointed out by several people who have
abandoned the exhaustive registration of large corpora,
claiming that the books are all alike in content and attitudes.
(See pages 272 and 380.)
  One problem with the usual type of registrations that
include more than alphabetical lists of titles is restricting the
number of classification categories. There are surprisingly
few examples of the use of text or pages from the textbooks
themselves to clarify or justify the categorization, such as
Mogens Jansen did (Jansen 1969). In his presentation,
facsimiles of textbook pages are not chance illustrations, but
part of the registration.
  Another example of how material can be physically
integrated into a study/registration is the large composite
volume Politik im Schulbuch (Pöggeler 1985), in which nine
different researchers analyze primers, religion books,
arithmetic books, biology books, song books and history
books. Both the titles and the descriptions in all the articles
are arranged in historical periods: Empire (1871-1918),
Weimar Republic (1918-1933), National Socialism (1933-
1945), Federal Republic of Germany (FRG) and German
Democratic Republic (GDR) (1945-1980). This is a tightly-
packed volume in which the average article length is 30
pages. Nonetheless, each article is accompanied by a large
number of pages consisting of facsimiles from the textbooks.
In addition, the bibliography provides accurate information
about the edition used, including the year of publication.
  No collective or systematic explanation has been given for
the selection of the reproduced primary material within the
parameter given (politics). The source is given: The books
come from Germany's third or fourth largest collection of
textbooks, the "Pöggeler Collection" in Aachen. Even though
some of the authors have also used other sources, the project
had to forgo any claim to comprehensiveness. The goal was
originally to carry out a "Gesamtuntersuchung" (collective
investigation) to be published in five volumes.
 
 
 

Perspectives
 
 

The survey above has mainly focused on studies that
register and treat a substantial number of textbook titles, of
which the majority in one way or another have been
incorporated into analyses, where they exist. The primary aim
of these investigations has been one of the following:

  - To achieve the most complete survey, diachronous or
    synchronous, of the stock of books, and perhaps in
    addition to draw up lines for further research (Jansen
    1969, Frische 1977, Choppin 1989),
  - to give a general survey of a school subject's content and
    development over a longer period of time (Tarschys
    1955, Andolf 1972, Møller 1983, Michael 1987),
  - to give a general survey of an academic discipline's
    content and development over a longer period of time
    (Bjørkvold - Hertzberg 1976, Chervel 1977),
  - to map the specific characteristics and development of
    a particular textbook genre, especially readers (Tischer
    1969, Helmers 1970),
  - to elucidate specific aspects of textbook content in
    accordance with paramount goals of a political nature
    (Georg Eckert Institute 1951, Pöggeler 1985),
  - to evaluate all or the greatest possible number of books
    in the existing stock of textbooks in light of general,
    overall pedagogical-didactic goals (Bjørndal 1982),
  - to elucidate the textbook situation in light of overall
    goals of educational politics (Sweden: Ds U 1978-1979,
    1980; Ds 1988).

The problem of method in connection with registration has
first and foremost been of a practical-economic nature. It has
been a question of time, resources and availability of sources.
As a rule, all registrations will of necessity be marked by
limitations in one or more of these areas. No matter how
methodically one works or how bountiful the resources at
hand, a shortage of source material will set limits. This is true
not only of historical investigations, but also of contemporary
surveys made since textbooks came into mass production. In
addition the problem of response represents a major
difficulty in most studies.
  The objective of the study must dictate the stringency of
the demands for comprehensiveness and accuracy in listing
authors' names, publication dates, book types, impressions,
 editions or sales figures. Thus the problem of registration
becomes at least as much a question of selection and reasons
for the selection, based on previous registrations, if any, as of
the systematization of known and new-found titles.
  Insofar as the relationship between objective on the one
hand and thoroughness of registration on the other is
concerned, the survey reveals great variation in solutions
from author to author. Some examples demonstrate that this
variety is not necessarily determined by differences in
approach at all. At one extreme we find Tingsten 1969, who
draws very general conclusions about nationalism in European
school books without attempting to disguise that his sample
is random. Lars Furuland has written an article in which he
points out that Tingsten's inaccuracy goes extremely far in
places:

  One important source is given as the Reader for Primary
  School; no more exact references are given. In the first part
  of the chapter "the line of religious propaganda" is
  illustrated with striking examples of children's beliefs
  about God, not least about fear of punishment from the
  hand of God. Only one of the ten or so quoted texts is
  taken, as nearly as I could find out, from Carlson's reader
  (editions 1-7). (Furuland 1988, p. 8.)

The method of registration and the specification of titles is
also limited in Bjørndal 1982. As with Tingsten, however,
this is not necessarily a mistake; the author emphasizes that
the work is chiefly a subjective interpretation undertaken
with the intention of helping to improve the books. The
material was meant to provide a foundation for discussing the
didactic condition of the entire book stock (see page 275).
The situation is nevertheless different from Tingsten's case.
Bjørndal's report is written on the basis of a school research
project carried out in an educational research center. One
hundred and thirty-three books were analyzed without any
explanation of their selection and without the inclusion of
quotations or specific references. In other words, the system
is the exact opposite of that which Elson 1969 used in the
examination of 1000 titles where every quotation and
reference was identified by edition and page. This is
unacceptable based on standard practice and the
methodological requirements applied to such studies.
However, as an experiment and an attempt to go in new
directions, it is interesting. It looks as though the researchers
dared to take the leap into subjective approximation because
their goal was so patently practical. If so, this is a dubious
justification at best. In one of the subjects, Norwegian, the
selection was limited to readers. But to what extent can the
study's point of view about readers' didactic function benefit
textbooks in Norwegian as long as Norwegian is a subject in
which the inherent integration of language and literature is
one of the primary goals?
  The modest degree of registration in Tarschys 1955 is a
consequence of the author's placing more emphasis on sources
 other than textbooks for the history of mother tongue
instruction. Even though the subject matter and didactic
viewpoints as expressed in the books receive considerable
attention, the author does not explain her sample, nor does
she address the question of how representative it is. Textbook
passages are brought in as illustrations, especially of
(diverging) professional opinions. They become symptomatic
of attitudes more than expressions of the tenor of the schools.
With the greatest respect for the breadth of documentation as
regards shifting professional views, it might be claimed that
Tarschys' attitude toward the registration of textbooks - in an
historical presentation of mother tongue instruction - is
nonchalant insofar it says nothing about selection or
priorities. Furthermore, the thesis was written in 1955, i.e.,
at a juncture in time when the stock of textbooks was
undoubtedly far more surveyable than ten or fifteen years
later.
  A jump from the studies mentioned above to those which
come closest to exhaustiveness also reveals different patterns.
Such virtual exhaustiveness can be achieved at the expense of
something else; it can take time and attention away from the
central issue. It may boil down to a question of whether
meticulous registration work is always equally necessary;
whether it serves a purpose, beyond its use as a scholarly
"alibi". In Frische 1977, the author points out that the
registration of the primary works used in 1,470 graduating
classes is unnecessarily lengthy in relation to the analysis, and
explains the discrepancy quite simply as the need for an
overview (p. 175).
  In his historical analysis of the development of the subject
of mother-tongue instruction in Great Britain (Michael 1987),
Ian Michael used the extensive registration in itself as an
argument for the hypothesis, i.e., that such a widely taught
discipline did exist. With Bjørkvold and Hertzberg 1976, the
correspondence between registration and selection was also
methodically functional. In spite of this, there is, as the
authors also point out, an element of self-contradiction in the
delimitation. It is expressed in the title (School Grammars);
how does one distinguish between subject content and
educational content in classifying titles that do not state any
specific target group?
  Andolf 1972 emphasizes the same sources as Tarschys
(public documents, polemical literature, scholarly literature).
The difference is that Andolf uses textbooks in addition as
his primary source. Painstakingly accurate registration work
thus becomes one of the pre-conditions for delivering what
the title (Teaching) promises.
  Kepner 1935 stands apart. Not only does he emphasize the
conclusions that can be drawn from an analysis of books that
were not widely distributed, in clear contradiction to Olsson
1986, for example. By attaching special importance to books
 written by authors established in other contexts as well, he
also differentiates between types of author. This broadens his
perspective and has probably had an effect on his results (see
page 355).
  The possibilities for comprehensiveness increase when the
registering is contemporary, at any rate if the area/country is
not too large. Jansen 1969 and Ds 1988 are examples of two
types of virtually complete surveys of book stocks at a given
time, motivated by research and school policy considerations
respectively. In the first case the survey was virtually
exhaustive because several groups were able to work in
parallel and thereby serve as reciprocal controls. In the other
case the guarantee lay in the mathematical-statistical methods
and a large government polling system.
  Like other literature, textbooks have genre characteristics
that can vary synchronously and diachronously. It almost
seems an unwritten rule that such characteristics not form
part of the registration. The exceptions are Göran Andolf,
Staffan Björck, Mogens Jansen and Ian Michael, who all
included some information about genre. The omission of this
aspect must be considered open to discussion. A
demonstration of diversification and variation - or of one-
sidedness - will be of importance, among other things, for the
discussion of the view that all the books are "the same".
  One recurring problem is the question of the possibility of
and methods for a more global registration than any
undertaken to date. In 1990 the National Psychological-
Educational Library in Sweden published a report on the
availability of textbooks from other countries as the
foundation for research (Hansson 1990). The report states that
the library has over 100,000 textbooks, of which
approximately 25,000 are foreign. The report concludes with
a proposal for a permanent arrangement for the purchase of
foreign textbooks in core subjects for an annual sum of
approximately 100,000 Swedish crowns. No comparable
arrangement has been either established or proposed in the
other Nordic countries.
  The fact that certain countries (such as France; see page
358) have come far, particularly in the field of historical
studies, therefore supports the contention that this must have
at least as much to do with the assignment of cultural policy
priorities and strong literary traditions as with methodological
inroads. Granted, the technological developments of the past
few decades have expanded the choice of registration
methods. But other countries with the same access to
electronics have not availed themselves of the possibilities to
the same degree.
  Theoretically, the possibility for computerized registration
should enhance our chances of gaining insight into our own
times. Such endeavors are taking place not only nationally,
but internationally. Pivotal institutions are The European
Documentation and Information System for Education (the
EUDISED database) at the Council of Europe in Strasbourg,
the Georg Eckert Institute in Braunschweig, the National
 Institute for Educational Research in Paris, and the
International Educational Library in Budapest. However, the
two-fold problem remains the same. We lack the capacity and
routines that would automatically bring new editions into the
system and make the material known and conveniently
available to those who need it. The Emmanuelle system in
Paris has probably come furthest. But even when technical
conditions are favorable, language and media policy
complications cause problems for the transfer of systems
across national borders. The ideal, total survey based on
continuous, automatic exchanges of information is more a
problem of cultural policy and communication than one of
machine capacity.
 
 
 

Go to: | Top of page | Next >