A Stereotyped Minority "Zigeuner" in Two Centuries of German Reference Literature

This paper reviews the entry "Zigeuner" in thirty editions of general encyclopae dias in the Germa n language between 1819 and 1986. The entries were found to be largely repetitive, stereotyped, and much of the "information" plainly fictive. While the homogeneity of "Gypsies" is concluded explicitly in almost all these editions, there is a chronological variation in the stereotypes. The entries exhibit both racism and romanticism, both fear and fascination. On th e whole it seems that what are supposed to be facts about "Gypsies" were and sWl are more often than not facts about their social environments.

The term "stereotype" is often used in common parlance, and surprisingly enough even in more specialised publications as if it were something static.But stereotypes changethey have a past, a present and a future.Like a prejudice, a stereotype " ... is a product of situations ... " (Schermerhorn 1978: 6), and as imperial discourse reveals, stereotypes can be switched to suit the convenience of those in power (cf.Gordon 1992: 213).Often, even within a short time span subtle changes may take place in a stereotype, although the corresponding changes in real life probably took place earlier (Schweizer 1980 : 40-41;Rao 1986Rao , 1988)).However, even over long periods of time the broad outlines, and even more so the ideological contents of a stereotype tend to vary only slightly, if at all.This applies to both auto-stereotypes and hetero-stereotypes, to both positive and negative ones.Simply expressed, a stereotype is a relatively standardized and fairly widespread image about a group considered distinct and fairly homogenous among this and/or any other group.Since it is standardized and popular, it has to be simplified; nuances have little place in a stereotype .This paper will endeavour to provide a review of certain relatively standardized pieces of printed information on "Zigeuner" ("Gypsies") for the general reader in the Germanspeaking world.Is this information stereotyped?If so, in what respect, and to what extent is there any change in it over the years?The data analysed are drawn from thirty encyclopaedias; although they are not exhaustive for the period examined, we believe them to be representative of their genre.Out of our sample thirteen were published in the 19th century and sixteen are from the 20th century.The nineteenth century_ was an extremely fecund period for more specialized publications on "Gypsies" in German, and the influence of some of these works is evident on the entries examined.A very competent analysis of the "Gypsy"-image in semi-popular journalistic publications of 19th century has been provided by Hehemann (1987: Ch. 3.).Willems and Lucassen (1990) have done a similar and interesting study on Dutch publications, while Mayall (1992) provides a review of the topic for Britain.For reasons too well known to discuss here there have been relatively few specialized publications in German in this century, and especially after the Holocaust.The press has been largely prejudiced and hostile even after the second World War, and as Franck (1987: 29) notes, there has been no change in " ... die ungebrochenen Vorurteile, die bis in die siebziger Jahre hinein auch in Schulbiichern zu finden waren und denen bis heute in Schulbiichern keine Aufklarung entgegengesetzt wird ... ".
Jochimsen's survey concerning prejudice in Germany (Vossen 1983: 131-133) also shows clearly that "Gypsies" head the list of discriminated groups.
It is hoped that the present study will also be a useful supplement to Gronemeyer's (1987) interesting commentaries on the various publications which appeared prior to Grellmann's work.

Encyclopaedic information and stereotypes
After the great encyclopaedists of the 18th century had launched their idea of compiling various compendia of knowledge available to them at the time, gradually a number of such comprehensive works on specific subjects were published in different parts of Europe, and over the years their numbers grew.To begin with, the readership aimed at was restricted to the elite few who had the benefit of a general education, and additionally deeper knowledge in say, jurisprudence or theology.But with time the level of general education rose, the target public expanded, and so did the volume of generalized and specialized knowledge.By the 18th century the purpose of an encyclopaedia was to present its readers with something like a brief"state of the art" on each entry.One of the first such general encyclopaedias in the German language appeared in 1721, and was entitled "Allgemeines Lexicon der Kilnste und Wissenschafften".Its author and publisher, Thomas Fritschen, devoted 467 words to an entry reading "Zigeuner, Cingarus, Bohemien, Egiptien".Twentyeight years later in 17 49 we find an entry of 5600 words on "Zigeuner" in another of the early German encyclopaedias.

112
This "Grosses Vollstiindiges Universal-Lexikon Aller Wissenschafften und Kilnste" was compiled and published by Johann Heinrich Zedler.As its first page indicates, it aimed at covering all the human arts and sciences including all major geographical, political, historical, genealogical, religious, mythological, and educational aspects.The latest encyclopaedia of general knowledge to appear in German is the preliminary and abridged "Vorauslexikon Brockhaus Encyklopaedie" of 1986.
Encyclopaedias of general knowledge are considered comprehensive reference works; they are meant for the general public and are expected to have a fairly wide distribution.The individual entries must hence be of a somewhat standardized and simplified nature.Thus, it seems appropriate to ask in what measure the information available in such encyclopaedias under the heading "Gypsy" are stereotyped, what is their content, and what is the type of image projected (cf.Willems and Lucassen 1990).

Problems of categorization and analysis
Entries regarding "Zigeuner" and "Ziegeuner" in thirty editions of encyclopaedias in the German language were analysed; of these fifteen editions were of the Brockhaus, seven of Meyer, four of Herder, and one each of Bruere, Duden, Knaur and Pierer (see Table I).The information in all these sources was of an almost entirely textual nature; thus data analy- (Knaur 1975) and 20 (Meyer 1952) pages of different sizes it was decided to count the number of statements and words in each entry.The number of references for further reading was also counted, as an indicator of the interest the reader was expected to have in the subject (cf.Table I).
For the qualitative analysis of the texts, all the words and/or statements in each entry were categorized as "neutral", "positive ", "negative", or "ambivalent".The category "neutral" consisted mostly of statements which could, at least theoretically, be fairly easily verified.By and large no adjective (e.g."geheimnisvoll" = "mysterious") was included in this category.
In categorizing these textual contents we had to try and take care not to fall into the pitfalls of chronocentrism, or what Banton (1983: 33) terms "presentism", i.e. the interpretation of other historical periods in terms of the values of our time."The historical study of racial thought and attitudes has often been flawed by an unreflecting presentism ... ", writes Banton, and it was partly in our endeavour to avoid such flaws that the category "ambivalent" became important .On many occasions in case of doubt it was considered more prudent to categorize a statement, a substantive or an adjective as "ambivalent" rather than as "neutral", "positive" or "negative".Within the category "neutral" a further distinction was made between what we term "active" and "passive" statements; the latter consist entirely of statements regarding the treatment meted out by the Gaje (i.e.non-Gypsies including "Gypsologists") to "Gypsies", and at times include short commentaries on this behaviour.The absence or presence of such commentaries are also interesting indicators of the writers' attitudes towards "Gypsies".Table II presents typical examples of terms and statements categorized by us as "neutral (active or passive)", "positive", "negative", or "a1nbivalent".

A categorization of the entries
As mentioned earlier, the entries examined vary in length between seventy words (Duden 1968) and 19040 (Meyer 1852).The earliest entry (Brockhaus 1819) has 2150 words and the last unabridged entry (Brockhaus 1981) has 2184 words.The length of an entry depends on two factors.Firstly, to a certain extent, on the entire length (i.e. the total number 8* of volumes) of a given encyclopaedia edition (see Table I).The length of an entry in an edition with four to ten volumes was never found to exceed 2150 words.The length of entries in editions with twelve to twentysix volumes, however, varies greatly and lies between 486 (Meyer 1977) and 4544 (Brockhaus 1895) words; an extreme value of 19040 words was found for one entry in the fifteen-volume edition of Meyer (1852).The second factor on which the length of an entry apparently depends is its period of publication.Entries of the 19th century tend to be longer than those of the 20th century.Without entering into details it could be suggested that the longer entries in the 19th century reflect the "romantic" interest in everything "exotic".
As can be expected , the longer an entry the greater the amount of information.Three questions now arise: first, what manner of information is given, and is there temporal variation in this?Secondly, to what extent is the information in the various entries identical, and if there are differences what is there nature?Finally, has this information been found to be correct, i.e. objectifiable at least at the time of writing?
The type of information given in any entry is implicitly or explicitly classified as under the broad headings "history, "demography and geographical distribution", "language", "habits and customs, including religious ones", "physical characteristics".The proportion of information under each of these headings was found to vary somewhat over time.For example, whereas Brockhaus (1819 and 1822) have only a couple of sentences each regarding language (2.5% and 4.0% of their words respectively), some years later Pierer (1836) devotes almost half a page (13.3% of his words) to it; this upward trend continued for a while, and although in 1845 Bruere has only 0.3% words on language, Meyer (1878) takes up about three-fourths of a page (83.3% of the entire entry) discussing aspects ofRomanes.Theimportance given to language (but cf.Willems and Lucassen 1990: 38) declines after a while, and in 1897 (Meyer) only 10.5% concern this topic; Meyer (1930 and1964) have 14.2% and 8.4% respectively.In more recent years the figures vary between individual encyclopaedias: 5.3% in Knaur (1975), an entire subsection (47 words) on "Zigeunersprache" in Meyer (1977), 2.7% words in Brockhaus (1974) and 2.5 % in Brockhaus (1981).
In quantitative terms, information supplied under the heading "physical characteristics" varies little over time, ranging between 0.0% and 6.3% in 25 (86.2%) of the entries analys ed.Exceptions are Brockhaus (1855) and Meyer (1878 ) for the 19th century (with 33.9% and 13.2% respectively) and Brockhaus (1957) and Meyer (1979) for the 20th century (with 9.0 % and 11.4% respectively).The variation in qualitative terms is of interest here.
It is only from 1822 (Brockhaus) onwards that every entry has something to say on this subject, and the terms classified by us as "positive " are often drawn from these sentences.Many sentences are repeated verbatim in various editions.Beautiful white teeth and finelyshaped limbs are invariably mentioned; skin , eyes and hair are often described in a manner that is either positive (e.g."hiichst lebendige Augen"), ambivalent (e.g ."blitzende Augen") or even negative (e.g."Aus den gluhenden Augen blitzt thierische Wildheit hervor; unstet schwankt der Ausdruck zwischen Schlauheit, Furcht und HaJL." -Meyer 1878: 1001).In the middle of the 19th century supposed physical traits started being correlated to alleged psychological characteristics; thus in Meyer (1852 : 781) we read, " ... seine Physiognomie zeigt Leichtsinn und Gemuthlichkeit".In the 1890s the entries come gradually under the influence of studies on race and physical anthropology.Meyer (1897: 1025) states: "Ethnologisch sind die Zigeuner wohl als ein Mischvolk zu bezeichnen, das man nur mit Vorbehalt zu den Ariern rechnen darf."Herder (1907Herder ( : 1827) ) lays even more emphasis on racial aspects: " ... arisches Wandervolk ... vielfach mit Fremden vermischt, die mehr und mehr schwindende reinblutige Minderheit mittelgroB und daruber, wohlgebaut, olivenfarbig, mit langlichem Gesicht, gerader oder Adlernase ." Whereas Brockhaus (1920: 972) compares them to the Dards, Meyer (1930Meyer ( : 1788) ) opines that they are " ... nur mit Vorbehalt zu den Ariern zu rechnen; sie stehen der nichtarischen Urbevolkerung Indiens nahe."Till the 1930s, in keeping with sexist Gaje attitudes (Rao 1985), "Gypsy" women were invariably described as being very beautiful; in Nazi Germany the 1935 edition of Brockhaus (p.630) finds a way out of its political dilemma by simply stating that the proverbial beauty of the "Gypsy' ' woman fades very early (cf.Willems and Lucassen 1990: 32).It is also in 1935 that we read for the first time (Herder, p. 1506) that "Gypsies" everywhere are racially very different from the surrounding populations; Brockhaus (1938) publishes the first portrait photographs with captions on colour of hair and eyes -taken from Nazi records -and mentions that in Germany "Gypsies" are covered by the "Blutschutzgesetz".
More absurd still in respect of physical characteristics are the entries after the second world war.From 1957 onwards all Brockhaus editions state that "Gypsies" are "related" to Dravidian and Mon-Khmer peoples, and till 1986 all entries (including those published in the German Democratic Republic) speak of mixture of race and blood .
If we take a look at our "passive" sub-category of "neutral" statements, we find several examples of sentences dealing with the persecution of "Gypsies" by the Gaje.Brockhaus (1819: 754) speaks of the stringent punishments imposed on them, but tries and justifies these: "Als die Beschwerden uber sie so haufig und laut wurden, fing man an, sie strenger zu behandeln; aber die hartesten Strafen fruchteten nicht gegen sie.Sie wurden daher nach und nach aus den meisten Staaten verwiesen, und als sie sich dennoch hin und wieder in kleineren Haufen einschlichen, wurden gescharfte Verordnungen gegen sie gegeben." The author goes on, however, to explain their nomadism, or rather what he calls their "herumschweifende Lebensart" ("vagrant lifestyle") as a result of these policies .Further, he writes that the Empress Maria Theresia had tried to turn "Gypsies" into "useful members of the State" (" ... brauchbare Mitglieder des Staats ... "), but inspite of all these attempts "most of the population continued to practice their old uncouth customs" (" ... blieb der gToBte Theil dieses Volks noch bei seinen alten rohen Sitten").Should one then assume that these "savage customs" were the result of Government policies which forced "Gypsies" into "vagrancy" -the contradiction in the author's statements were apparently not obvious to him.The next edition of the Brockhaus (1822: 898) was even more straightforward in its commentary of "Gypsy" persecution: "Doch halfen selbst die Verfolgungen nur auf kurze Zeit." ("Even the persecutions helped only for a short while").Without attempting any explanation of their nomadism, as his predecessors had done he describes briefly how the "great Maria Theresia tried to make human beings and citizens" ("Menschen und Burger") out of "Gypsies".Eighteen years later, Pierer is categorical in his opening sentence that the "herumschweifende(n) Zigeuner" are without "eigentliches Gewerbe" ("real profession").At the close of the 19th century we find a slightly different tone in Meyer (1897: 1024): "Die Zigeuner . . .wurden . . .infolge ihrer Betrugereien und Diebstii.hlebald auf das grausamste verfolgt, ohne daB man jedoch das unheimliche Volk auszurotten vermochte ... aber Mari.aTheresia und Joseph II. (erzielten) durch menschliche Behandlung ... der Unterdruckten gute Erfolge ... ".
No explanation is given for their itinerancy, although it is referred to.The Brockhaus of 1920 (p. 973) represents a further development in this change of tone: "Die Geschichte der Zigeuner ist eine Geschichte menschlichen Elends und menschlicher Rohheit.Zahllos sind die Edikte, die in allen Herren Lander gegen sie erlassen worden sind, und grausam die Verfolgungen, denen sie ausgesetzt waren." Here nomadism is referred to but no reasons are given.It is surpnsmg that all but one (Meyer 1979) of the entries fail to mention the economic relevance of "Gypsy" nomadism.In the light of what Gunther (1985: 12)  In fact, although some entries mention "Gypsy" slavery, few of the even recent entries mention the fact that a large part of the population has been sedentarized for long.In a society such as that of Europe where sedentism has had such an ideological value, the label of a nomadic life-style sticks on for a long while: "The Gypsy image has evolved over time from that of 'adventure seekers' and highway bandits to vagabonds, then to nomads and then again to 'people without fixed abode', before ending up as 'people without fixed and regular income enabling them to lead a normal life'.This last image has in its turn, now become one of 'itinerant people', or 'people of nomadic origin'" (Liegeois 1987: 363).
In the 1930s no mention is made of any previous (or contemporary) persecution, and in the post-war period Herder (1956Herder ( : 1442) ) goes as far as to explain that the Nazi regime persecuted, deported and exterminated "Gypsies" because of the problem of their integration as a nomadic population among the sedentary culture of their hosts.Their 1968 edition still speaks of this problem of integration; but no persecution, let alone the Holocaust is referred to.The same is the case with Duden (1968).Knaur (1975) perhaps found words such as "persecution" and "genocide" too strong and chose instead to write that throughout their history "Gypsies" were discriminated against and did not enjoy full rights ("minderen Rechts").But then it is suggested that perhaps all this did lead to something positive -namely "the survival of the most important aspects of their language and culture"!Brockhaus (1957) does not spare a single word to even refer to any persecution at any time; their next edition (1974), somewhat like Knaur (1975) avoids all "provocative" terms.Certain information is, however, listed in a very matter of fact manner.The identical entry appears in the edition of 1981, and the 1986 (preliminary) edition makes minor but important lexical changes for the Nazi period.Instead of "durch den Nationalsozialismus hatten die Zigeuner sch were Verluste (... die Zahl en schwanken zwischen 250 000 und 500 000).Vom 16.12.1942an wurden alle Rom-Zigeuner und Mischlinge nach Auschwitz gebracht ... ", we now read: "Wahrend der nationalsozialistischen Herrschaft wurden etwa 500 000 Zigeuner in Konzentrationslagern getotet (1982 offiziell als Viilkermord anerkannt)." Finally, the last two editions of Meyer (1977 and1979), both published in the German Democratic Republic, mention the never ending persecutions, but give details only of the genocide in the Nazi period.The 1977 edition also implies that after the creation of the "Peoples' Democracies" in eastern Europe and the sed-118 entarization there of most "Gypsies", discrimination and persecution came to an end.Contemporary racism and discrimination in Rumania, Poland and erstwhile Czechoslovakia as well as the incidents at Rostock belie this information.
Our second question, concerning the similarities and differences in the entries, has been briefly touched upon in the above examples.At certain periods different editions of the same encyclopaedia contain almost identical texts.Thus, whereas the texts in Brockhaus (1819Brockhaus ( , 1822Brockhaus ( and 1841) ) are fairly distinct from one another, Brockhaus (1848) takes over large chunks of the Brockhaus (1841) entry; the entry in Brockhaus (1920) is a reprint of the 1908 edition.Similarly Brockhaus (1974) and Brockhaus (1981) have almost identical entries; during the same period Meyer (1977 and1979), however, have entries which are very different from each other.
We now move on to the question of the veracity of the information given.Although we shall now primarily consider examples of statements which we have categorized as "neutral", since these are the most easily verifiable, the other categories are inevitably touched upon in the longer examples discussed.The earlier entries had little source material to draw upon, and their statements are often plainly fictive.A case in point regards the choice of marriage partners where many entries speak of incestuous ties.Brockhaus (1855: 525) writes generally of marriage between "blood relations" ("Blutsverwandtschaft"), while ten years earlier Bruere (1845: 1200) and even some twenty years before Pierer (1836: 663) are more explicit in stating that when a boy reaches puberty he chooses a wife "irrespective of whether she is his sister or any other near relative, or a stranger."( ... unbekiimmert darum ob das Madchen seine Schwester oder eine Fremde ist..." and "unbekiimmert ob sie seine Schwester oder sonst eine nahe Verwandte ist ... " in Bruere and Pierer respectively).Even later, the unreliability of the source material renders most information in these entries rather questionable.Especially as far as ethnographic information is concerned, the entries do not by and large even refer to the contradictions in the very source materials they use; Brockhaus (1895: 973) is an exception in that it gives an annotated bibliography and warns the reader about the reliability of certain publications by Wlislocki.The result is that all statements of all "Gypsologists" get treated more or less as Gospel Truth; a pattern which had tragic consequences in the Nazi period.Martins-Heu£ (1983: 59) sums up the problem neatly: "Aus der Unfahigkeit der Nichteingeweihten, die Qualitat der sich haufig widersprechenden Aussagen der Forscher zu differenzieren, resultierten der Effect und die Unantastbarkeit der Zigeunerforschung und ihrer Vertreter.Die interessierte Offentlichkeit sowie das Gros der Fachkollegen waren gewillt, alles zu glauben." In this connection it is interesting to briefly review six standard readers of ethnography and ethnology published in German between 1830 and 1900.Each of these has a short section on the "Zigeuner".All draw implicitly or explicitly on the studies of"Gypsologists", such as Grellmann, Heister, Miklosich, Stumpf and Trumpp.For the Volkergalerie (1830-39) the "Gypsy" is a barbarian, depraved and degenerated, but he is in a way absolved, since "nur das Leben verwandelt diese Menschen und theilt ihnen ihre ganze Widrigkeit mit ... " (p.149).Even this meagre "sympathy" is lacking in Gustav Klemm (1851), who devotes 1710 words to describe this "active race" (p.284) who are nevertheless "the worst enemies of all work" (p.285), "sensuous, cruel and without a trace of the sense of honour" (p.287).In his subchapter on the "Aryan Peoples of Asia", Perty (1859) includes a paragraph (pp.93-94) on "Gypsies"; his information is of an historical nature and deals almost exclusively with Asia.Peschel ( 1875) is even more brief and dismisses the "puzzling Gypsies" (p.541) in a single sentence in his section on the "Indoeuropean Branch" of languages.A very similar entry is found in Muller (1879) in his section on the "Indogermanic Branch" (subsection: "indic family").The last reader examimed, Reymond's "Illustrierte Lander-und Volkerkunde" appeared at the turn of the century and includes a picture of "Zigeunerinnen" and a passage of350 words on "Gypsies" in its section on Austria and Hungary (p. 383).The tone of the passage ranges from patronizing to romantic.

Image(s) projected
In his by now standard reader on social psychology Hofstatter (1963: 391) affirmed, that "Die Rolle, die eine bestimmte Minoritat in einer bestimmten Gesellschaftsordnung spielt, fuhrt zu Stereotypen, in denen sich die bestehenden Spannungen widerspiegeln." This takes us back to our starting point and ask whether the information in these various encyclopaedias is stereotyped?If so, does the temporal change of information mentioned in the above paragraphs impinge in any manner on such an eventually stereotyped image?
A stereotype can only be about a group considered distinct and fairly homogenous by members of this group and/or others.The very fact that the distinct entry "Zigeuner" (and additionally sometimes "Zigeunermusik" and "Zigeunersprache") exists does suggest that the encyclopaedists did and do hold that this term designates a distinct population (cf.Lucassen 1991).Although Brockhaus (1895: 973) suggests that it is increasingly difficult to distinguish this population from those surrounding them (" ... sie gehen immer mehr in der Beviilkerung auf, in der sie leben"), Meyer (1977) states clearly that they constitute a " ... Volk, das als selbstandige ethnische Einheit in kleinen Gruppen von einigen Familien unter fremden Viilkern lebt." As regards the homogeneity of the group or the lack ofit some entries give somewhat confused but long lists of subgroups (e.g.Pierer 1836, Brockhaus 1974and 1981); language and other regional differences as well as "racial mixture" are also mentioned in many entries, but the general style of them all points to a certain degree of homogeneity.Incidentally, attempts at tracing the term "Zigeuner" from

B % POSITIVE
Figur e 1.The variation in the percentag e valu es of statements categorized as either "positive" or "negative" in seven editions of Meyer and fourteen editions of Brockhaus.
"Zieh-gauner"an etymology which among other things could also imply a certain heterogeneity of the population -are mentioned for roughly fifty years from 1822, but are uniformly rejected by the authors.In fact the homogeneity of the group is explicitly concluded by referring to it in almost all entries as one "Volk"; it is implicitly done by the repetition in each entry of a certain number of statements portrayed as facts.The repetition of these "facts", no doubt somewhat con variationi over the centuries, leads to a certain standardization and over-simplification.It is also this bundle of "facts" which constitutes in the final count the stereotyped "image" projected by, and in, these various entries .
"Versions will differ in the amount of homogeneity they impose internally; and some may permit great internal discontinuity of texture, style, or perspective" (Gellner 1987: 166).As is well known, neither romanticism nor the influence of Rousseau however, prevented or even diminished the persecution of "Gypsies", and in 18th century Prussia the persecution of Sin ti reached new heights (Spaich 1981 : 19).Generally in Europe, from the late 18th century onwards culture-difference came to be increasingly classified as cultural and indeed overall retardation.As Gellner (1987: 47) expresses it, " ... backwardness replaced evil as the generic characterization of that which is undesirable." But following the biological and social think- -% POSITIVE ing of the time (Engels, Haeckel, Levy-Bruhl, etc.) there were degrees even in backwardness, as Krauss (1987: 103)  The chronology of stereotypes Many of the entries examined suggest, to say the least, that "Gypsies" are evil -they have no real religion, their sexual morals are highly questionable, they are thieves and rogues, and according to Brockhaus (1822), Pierer (1836) and Bruere (1845) they take pleasure in being cruel.No wonder then, that as in the case of all evil they are as "fascinating as terrifying and repellant" (Brockhaus 1855: 523).The final act of evilness is their "rejection of the culture of .their host societies " (Brockhaus 1957) .But all entries further project the image of an extremely "backward" people .The degree of "backwardn ess" appears to swing back and forth somewh ere between the above mentioned steps ofprimitivity and barbarity .Their "primitivity" binds them to nature: their "savage" life style (Brockhaus 1819), their "animallike love for children" (Brockhaus 1822), their oft commented on (cf Hancock 1992 : 11) "naked children " (Pierer 1836), their "wild looks " (Bruere 1845), the "open sky which serves them as shelter" (Brockhaus 1868), their "sup erstitious fear of the dead" (Meyer 1930), "their old heathen customs" (Herder 1935) all add up to their being classified as a "Naturvolk " in 1935 (Brockhaus,p. 630) and their way of life being clearly stamped in 1957 (Brockhaus,p . 698) as "very primitive" with "traces of matriarchal concepts" (Herder 1956).Their barb arity is on the other hand equally obvious from their "thoughtless, faith-less, fearful" (Meyer 1878), and "merry, carefree, timid and suspicious" (Herder 1907) nature.Strange, that in the midst of so much primitivity, irrationality and childishness Brockhaus (1895) could state: "Die Zigeuner haben eine eigene politische Verfassung" -in other words a complex system of making decisions and translating these communally authorized decisions into planned action.On the whole, however, this "mysterious" (Meyer 1852) and "strange" (Brockhaus 1895) people, without a land to call their own ("Heimat": Brockhaus 1841), but with a "king" of their own (Brockhaus 1848: 550) are both evil and backward, both primitive and barbaric: "Ihrem Character liegt Liebe zur Freiheit oder eigentlich zur Ungebundenheit, dabei zum Betrug und Diebstahl, so wie Triigheit zu Grunde" (Pierer 1836: 662).
An examination of the statements categorized by us as "positive", "negative", or "ambivalent" over the roughly one hundred and fifty years further suggests that there was a downward trend in both "positive" and "negative" statements, and fairly constant values below 5% (except for Bruere 1845: 8.5%, Meyer 1852: 10.8%, Herder 1956: 11.5% andHerder 1968: 7.1 %) for ambivalent statements.Although all the entries showed this trend diachronically, it was decided to plot the values only for those encyclopaedias for which a sample of more than five editions was available.Figure 1 shows that the entries in Meyer had values between 11.8% (1878) and 1. 1% (1977) for the "positive" statements and between 12.9% (1878) and 3.5% (1979) for the "negative" ones.Although the values for Brockhaus show a similar trend, the "negative" values in the 19th century are somewhat higher than in Meyer; in the Nazi period (Brockhaus edition of 1938) we not surprisingly find a remarkably high value for "negative" statements, with a correspondingly low value for the "positive" ones.It is also interesting to note that, except for this 1938 edition, the "positive" and "negative" value-sets -especially in Meyer -run roughly parallel to each other.This parallelism reflects the unabashed subjectivity of the earlier au-122 thors, who took it for granted that their opinions, emotions and biases were there to be shared by their readers.This subjectivity incorporated both racism and romanticism, both fear and fascination.As Brockhaus (1868: 746, 7 4 7) expresses it, this "romantic tribe" was also a kind of"weed which continued to grow".It was only in the Nazi period that racism and fear finally entirely overcame romanticism and fascination; the end products were hate and genocide.Thus, although the "negative" values are always higher than the "positive" ones, a certain "balance" is always achievedexcept during the Holocaust.
In his paper on the historical relationship between the development of ethnography and folklore studies (Volkskunde) in Germany, Kiinenkamp (1988: 32) writes that encyclopaedias provided a kind of academic guidance.
"Man durfte also hoffen, Schriften dieses Genres verbindliche Auskiinfte uber systematischen Ort und Inhalt der Ethnographie zu entnehmen ... " In each of the entries we did indeed, find "facts".The question is: were these "facts" about "Gypsies" or were they rather, overwhelmingly about the social atmosphere of the time?The various visions presented here, we suggest, were related both to the manner in which the vision of these groups changed and to the change in the culture of the viewer (cf.Willems and Lucassen 1990).A diachronic examination of the entries indicates, as mentioned above, that these "facts" are a series of reactions to similar, older "facts" as has so often been the case in "Gypsy" history (Gilsenbach 1985).The decrease over the years in the percentage of both "positive" and "negative" statements reflects an attempt to (appear) more objective."Facts" have been sifted down over the centuries to reach an embarassed "neutrality".By being so "neutral", the modern encyclopaedists would have us believe that simple, "unbiased" facts can not breed repression and rejection, and that the persecutor must be blatantly subjective.As with all stereotypes, here too the author's individual responsability for the views aired and the verac-

Table I .
Variation in the length of entries (i.e.number of words), the number of references cited in each entry analysed and the total number of volumes per edition examined.

Table II .
Examples of terms and statements classified for purposes of analyses in one of five categories.B = Brockhaus, Br = Bruere, H = Herder, M = Meyer, P = Pierer.