Loaded words: Evolving interpretations of ‘anti-semitic’ and ‘anti-semitism’ [1] in dictionary definitions and in public discourse [2]

Willem Meijs, Language Consultancy Desk, Birmingham, UK
Susan Blackwell, University of Birmingham

Abstract

Some words are loaded with connotative associations that make them highly sensitive elements in public discourse, especially political and legal discourse. This is certainly the case with the words anti-semitic and anti-semitism.

While Semites and semitic were originally used to refer to a broad ethnic category that included both Arabs and Jews, their derivatives anti-semitic and anti-semitism came to be applied, from first use, almost exclusively to people of Jewish ethnicity or religion, meaning roughly ‘hatred of / hostility towards Jews’. In some quarters over the past few decades there has been a further semantic shift, involving an extension of the meaning of anti-semitism to include criticism of, or hostility towards, the state of Israel. This paper traces these semantic shifts both in evolving dictionary definitions and in public discourse as evidenced in the Bank of English and the World-Wide Web.

More recently still, the European Monitoring Centre on Racism and Xenophobia (EUMC) devoted some emphasis in its 2004 report to the lack of a common definition of anti-semitism, and promptly offered one. The resulting “EUMC Working definition” has been taken up throughout the EU: for instance in the British Parliament through the report of its All-Party Parliamentary Group against Antisemitism (September 2006). This did not volunteer a definition of its own but concluded: “We recommend that the EUMC Working Definition of antisemitism is adopted and promoted by the Government and law enforcement agencies.”

Our study revealed that the terms semite, anti-semitic and anti-semitism are the focus of much linguistic contention, particularly in the UK. We found that collocational patterns in data culled from the World-Wide Web fluctuated widely from one year to another: disturbingly, this appeared to be largely due to the influence of pressure groups and documents in the public eye at the time. We conclude by drawing some salutary lessons for linguists and lexicographers.

1. Introduction

Our starting point for this paper is an article by the campaigning journalist Robert Fisk, published in The Independent newspaper in April 2004. In it he berates Webster's Third New International Dictionary for defining anti-Semitism as “opposition to Zionism: sympathy with opponents of the state of Israel”. Fisk goes on to quote “the pitiful response of the Webster's official publicist, Mr. Arthur Bicknell, who was asked to account for this grotesque definition”:

‘Our job’, he responded, ‘is to accurately reflect English as it is actually being used. We don't make judgement calls; we're not political.' Even more hysterically funny and revolting, he says that the dictionary's editors tabulate ‘citational evidence’ about anti-Semitism published in ‘carefully written prose-like books and magazines.' Preposterous as it is, this Janus-like remark is worthy of the hollowest of laughs. (Fisk 2004)

In this paper we will attempt to provide context to the conflicting claims by Fisk and Bicknell by charting the semantic shifts which have taken place in the use of anti-semitism, zionism and related words. We will examine the changing definitions in the most influential British and American dictionaries over the last century, and compare them with the use of the terms in current British and American English as found in newspaper texts in a general corpus and pages from online newspapers on the World-Wide Web.

This study bears similarities with the task which Baker et al. (2008) undertook: trying to trace and interpret the evolving use of certain ‘loaded words’ over time in the light of developments in ‘the real world’, utilising linguistic evidence as found in textual sources from relevant time-frames. Our analysis has been guided by Fairclough's Critical Discourse Analysis (CDA) (Fairclough 1995) and Critical Language Study (CLS) (Fairclough 1995, 2001). We have attempted to provide a “description” (Fairclough 2001: 21) of the linguistic characteristics of our chosen texts, using corpus linguistics methodology; an “interpretation” (ibid.) of the processes by which they are produced and interpreted; and also an “explanation” (Fairclough 2001: 22) of the wider social processes within which the semantic space of words like anti-semitism and anti-zionism is fought over.

2. Diachronic semantic shift in the meanings of anti-semitism, anti-zionism and related words

2.1 Biblical beginnings

Not surprisingly, both groups of words have Biblical origins.

According to Genesis 5:32, Noah fathered Shem, Ham, and Japheth at the somewhat ripe old age of over 500 years. After surviving the Flood, Noah, his three sons and their respective wives replenished the earth with human beings (Genesis 9:18). Japheth's descendants populated the coastland areas, while those of Ham included the Canaanites who spread from Sidon to Gaza, and Shem's descendants occupied territory which “extended from Mesha in the direction of Sephar, the hill country of the east” (Genesis 10:30). One of Shem's descendants was Abraham, who left Ur of the Chaldees in Mesopotamia to head for the land of Canaan (Genesis 11:31; 12:5): the rest is history, or at least mythology. The populations which allegedly derived from these families were labelled Semitic, Hamitic and Japhetic, after the three sons of Noah. Linguists will be more familiar with these terms as names for the language families associated with Asian, African and purported Mediterranean languages respectively.

The word Zion is the Old Testament name for Jerusalem, but even in Biblical times it was often used metaphorically or metonymically. Zion was the name of the hill fortress of the Jebusites on the southeastern hill of Jerusalem, which was conquered by David. It later became used to indicate the hill where David's Temple stood, containing the Ark of the Covenant; or the whole city of Jerusalem from which he ruled the united kingdoms of Israel and Judah (II Samuel 5:7). It is a loaded term, replete with connotations of longing for return to the Promised Land. Psalm 137 famously describes how the captive Hebrews wept by the rivers of Babylon when they remembered Zion.

2.2 More modern meanings

2.2.1 Semite, semitic, anti-semite and anti-semitic

The Oxford English Dictionary (OED) cites the Biblical origins discussed above, defining semite as “A person belonging to the race of mankind which includes most of the peoples mentioned in Gen. x. as descended from Shem son of Noah, as the Hebrews, Arabs, Assyrians, and Aramæans”. While the earliest recorded usage in this sense is from 1848, no indication is given in the OED that the usage is yet obsolete. (The OED gives as a secondary meaning “a person speaking a Semitic language as his native tongue”, although none of its cited instances illustrates this sense.)

However, when we turn to look at the adjective semitic, the OED states that “in recent use” it is often specifically used to mean “Jewish”. The same is said of the noun Semitism, where “in recent use” the term means “Jewish ideas or Jewish influence in politics and society”. “Recent use” here apparently means the 19th century, since instances are cited from 1885.

In the linguistic sense of the adjective semitic, however, the meaning is once again given as a wider one: “The distinctive epithet of that family of languages of which Hebrew, Aramæan, Arabic, Ethiopic, and ancient Assyrian, are the principal members.” The cited instances date back to 1813 but here again there is no suggestion that the usage is obsolete.

When we turn to the compound forms anti-Semitism, anti-Semite and anti-Semitic, the OED states unambiguously that these relate to a “theory, action, or practice directed against the Jews”. Again, the earliest citations are from the 1880s; none of the cited examples postdates the Second World War, and no indication is given that there has been any change in meaning.

In fact these terms are ideologically loaded and are disputed territory in current public discourse. To give an example of the nature of the controversy, consider this extract from an e-mail to the Vice-Chancellor of Leeds University from Waleed Kamal, former head of the Palestinian Solidarity Group at the University of Leeds:

I have noticed the following on your University Website: Lecture: Wednesday March 14, Hitler's Legacy: Islamic Antisemitism in the Middle East.

… ‘Islamic Antisemitism in the Middle East!!??’ Excuse me, but the middle east consists of Arab States, around 300 million Muslim Arabs whom you are accusing collectively by such a title.

I wonder what makes you speak of this? What would have happened if you were to speak about Hitler's Legacy and the Israeli Zionist Anti-Palestinians? Would this not sound much better? Would it not be more logical?

The term Anti-Semite is defined as Anti any descendant of Sam, (see Oxford dictionary), so how dare you speak of Islamic Antisemitism if Muslims and Arabs are Semites whether you wish or not.

(as quoted in the Jewish Chronicle, 16th March 2007)

As can be seen from our account above, Mr. Kamal is not quite citing the Oxford English Dictionary accurately. The OED does indeed define semite as a descendant of Shem; however, it defines anti-semite as “one who is hostile or opposed to the Jews”. Thus, while according to the OED Arabs are semites, it is not possible to be anti-semitic towards them. Mr. Kamal could be forgiven for being confused by this and for putting his own productive use of morphology into practice. Unfortunately the authors have no knowledge as to whether the Vice-Chancellor's reply exhibited a superior knowledge of etymology or semantics to that of Mr Kamal.

We now turn to Chambers (British) and Webster's (USA) dictionaries to see whether they can shed any further light on this point.

The first edition of Chambers (1901) gives Semitic as “pertaining to the Semites, supposed descendants of Shem, or their language, customs, &c”. It lists the Semitic languages as “Assyrian, Aramean, Hebrew, Phœnician, together with Arabic and Ethiopic” and helpfully adds that the term was “Applied by J.G. Eichhorn in 1817 to the closely allied peoples represented in Gen. x. as descended from Shem.”

The first edition of Webster's New International Dictionary, also dating from the turn of the century, is even more precise. Giving “One of the descendants of Shem” as the primary meaning, it adduces a second sense: “A member of a Caucasian race now chiefly represented by the Jews and Arabs, but in ancient times including also the Babylonians, Assyrians, Aramaeans, Phoenicians, and various other peoples of southwestern Asia.” Webster's, too, attributes the term to Eichhorn. It also adds a lengthy note on the supposed racial characteristics of semitic peoples which would certainly not be considered acceptable in a modern dictionary.

At this point in time there is a discrepancy between Chambers and Webster's in their definitions of the compound forms beginning with anti-. While Webster's gives a more inclusive definition of anti-Semitism as “Opposition to, or hatred of, Semites, esp. Jews”, Chambers appears to be rather dogmatic, acknowledging the term Anti-Semites only in the plural and defining them as “the modern opponents of the Jews in Russia”. It must be borne in mind that both dictionaries were published not only before both world wars but also before both Russian Revolutions (1905 and 1917).

By 1964, however, while Chambers' definition of Semite is very similar to that of the first edition, it has broadened its definition of anti-semite: “a hater of Semites, esp. Jews, or of their influence”. The definitions of both terms are retained in the 1988 edition.

Webster's, by contrast, has moved in the opposite direction. By the 3rd edition of the New International Dictionary, which first appeared in 1961, its definition of Semite has barely changed from that of the first edition. However, anti-semite and anti-semitism are now both defined only by reference to hostility to Jews. Moreover, a new second meaning of the term has appeared: “opposition to Zionism: sympathy with opponents of the state of Israel.” This is the sense about which Robert Fisk complained so vociferously.

In fact, Fisk was not the first to complain publicly. On 16th March 2004 the American-Arab Anti-Discrimination Committee (ADC) announced that ADC Communications Director Hussein Ibish had written to Merriam-Webster demanding that the company take steps to correct the definition of anti-Semitism in its ‘Third New International Dictionary', which had been reprinted in 2002.

According to linguist Geoffrey Nunberg writing in the New York Times:

a company spokesman defended the definition as ‘a relic’ based on a handful of citations from about 1950 in which anti-Semitism was ‘linked more or less strongly with opposition to Israel or to Zionism.’ He said that the sense wasn't supported by current usage, and added that it would probably be dropped when the company publishes a new unabridged version in a decade or so. (Nunberg 2004)

The response given by Merriam-Webster to the ADC – admitting their error and promising to correct it in due course – is at variance with the one that Fisk claims to have received, in which the company claimed “to accurately reflect English as it is actually being used” based on “citational evidence” published in “carefully written prose-like books and magazines.” We will have to await the fourth edition of Webster's New International Dictionary to see whether the definition is preserved or not.

Meanwhile we can only attempt to investigate for ourselves whether it can really be the case that the term anti-semitism has undergone semantic shift in opposite directions on the two sides of the Atlantic. If the dictionary entries are to be taken at face value, while anti-semitism in England originally only applied to Jews, and to Russian ones at that, by the 1960s it was applicable to other kinds of semitic peoples. Meanwhile, in American English, or at least in international English as seen from the USA, in the course of the same sixty years the term had moved from designating hostility towards semites in general, especially Jews, to designating hostility only to Jews and to the state of Israel.

We now turn to another thread in the web of words.

2.2.2 Zionism and Zionist

The terms Zionism and Zionist predate the state of Israel by more than half a century. Some ascribe the term Zionism to the Austrian publicist Nathan Birnbaum, who is said to have coined it in 1891; however, the term must be older than this since it appears in Webster's first edition (see below). The Zionist movement is generally held to have been founded by Theodor Herzl at the first Zionist Congress in Basle, Switzerland, in August 1897.

The first citations for both words in the OED are from the Jewish Chronicle in 1896. One of the first recorded usages, from 28th August of that year, is in the statement “Zionism does not necessarily aim at the formation of a Jewish State in Palestine.” This is reflected in contemporary dictionary definitions of the term, both British and American. The First Edition of Chambers's Twentieth Century Dictionary, published in 1901, defines Zionism as “a movement for securing national privileges and territory (esp. in Palestine) for the Jews”, while the First Edition of Webster's New International Dictionary, first published in 1890, defines it as “Among the modern Jews, a theory, plan, or movement for colonizing their own race in Palestine, the land of Zion, or, if that is impracticable, elsewhere, either for religious or nationalizing purposes”. Both these dictionaries, then, reflect the fact that, while there was a preference for Palestine on religious and historical grounds, it was by no means a foregone conclusion that this would be the location of the future Jewish state. Prior to the Balfour Declaration of 2nd November 1917, early Zionist leaders considered a variety of possible locations (see Herzl 1896): Leon Pinsker favoured Argentina, while at the Sixth Zionist Congress in 1903 Herzl proposed settlement in Uganda, which was then on offer from the British, and won a majority for the proposal but nearly split the meeting because of fierce opposition, especially from the Russian Jews present.

Both Chambers and Webster's later narrowed their definitions of Zionism. Chambers' Revised edition, 1964, does not differ from its 1901 definition; but Chambers’ English Dictionary, 1988, defines Zionism as “the movement which secured national privileges and territory in Palestine for the Jews and which now helps to maintain and develop the state of Israel”. Meanwhile Webster's 3rd New International Dictionary, which first appeared in 1961, gives as its definition of Zionism “a theory, plan, or movement for setting up a Jewish national or religious community in Palestine.” Not surprisingly, since the establishment of the state of Israel in 1948 both dictionaries have dropped any reference to locations other than Palestine as the target of Zionist aspirations.

2.3 Investigation

These evolving and at times conflicting definitions lead us to ask whether there is any evidence for a difference in modern usage between semite and anti-semite in the ethnic groups being designated; and whether the use of the latter term really has become extended in normal parlance to encroach on the semantic space occupied by anti-Zionist to denote critics of Israel, or whether this is merely wishful thinking, or even a susceptibility to Zionist influence, on the part of the compilers of Webster's Third.

In order to investigate these questions we conducted corpus analysis on the terms semite, semitic, anti-semite, anti-semitism, anti-semitic, zionism, zionist, anti-zionism and anti-zionist, examining both British and US usage in newspaper texts. We studied collocation patterns and concordance lines, looking for evidence of changes where possible. We now report our findings.

3. Corpus evidence for semantic shift in modern English

3.1 Bank of English

The first corpus which we mined for data was the 450 million word “Bank of English” (henceforth BoE), jointly owned by HarperCollins Publishers and the University of Birmingham. The BoE was a development of the Cobuild Corpus which had been collected as a resource for the compilation of the Cobuild dictionaries and language teaching materials. Texts were continually added to it over a period of some two decades, but by the time the corpus reached 450 million words in 2002 most of the constituent texts originated after 1990.

For British data, the subcorpora from the Times, Independent, Economist, Guardian, Sun and News of the World newspapers were combined, giving a total of 172,707,015 word tokens. These will be referred to in this section as the “British press”. For the US data, the pre-existing category “usnews” (10,002,620 word tokens) was selected, which will be referred to as the “US press”.

The software accompanying the BoE offers the facility to compute Mutual Information and t-score statistics for collocates of a given node word and to sort collocates in order of statistical significance. Since Mutual Information scores can produce spurious significance statistics for low-frequency words (because the expected frequency of co-occurrence will be predicted as less than 1, and so any actual co-occurrences will appear highly significant), we have only made use of t-scores here. However, these too are of limited usefulness for our purposes because wildcards are not permitted in the node word, so that it is not possible to group variant spellings of a term together. Moreover, hyphenated words are split and treated as separate types, so that it is not possible to calculate the most significant collocates of anti-semitism. The collocates of anti-semitism are treated as collocates of semitism, and anti itself appears as a collocate of semitism. When antisemitism is spelt without a hyphen it can be treated as a node word in its own right, but then the statistics will be skewed by the absence of the hyphenated version from the counts.

Because of these limitations, we have presented raw frequency figures as well as t-scores. The reader should bear in mind that these give no indication of relative frequencies between UK and US data, since the UK sub-corpus used was approximately 17 times as large as the US sub-corpus. We felt that, while this was far from optimal, it was preferable to restrict ourselves to data which had been collected by the same methodology than to supplement the US data with material which was not homogeneous and could not be searched using the Cobuild software. The authors do not suggest that any statistical significance can be inferred from the figures presented here; rather, our intention is to survey recent media usage of the ‘loaded words’ under examination, from both sides of the Atlantic. In fact, some of the most interesting features we found were unique occurrences because there are a number of unexpected and sometimes creative semantic patternings in the corpora. While of no statistical significance, we nonetheless believe that these ‘hapax legomena’ collocations illustrate the contested nature of the words under discussion.

When researching collocations we used a standard span of 4 words to the left or right of the node word.

3.1.1 Semite, semitic, anti-semite and anti-semitic

The term semite in the singular did not occur at all in the British press. There were 3 occurrences of semites, plural, one of which refers to archaeological evidence and not to modern peoples. The remaining two are clearly inclusive usages:

(1)

But antagonism between Palestine and Israel can hardly be racist, he says. ‘We are both the same race, both Semites. It is purely a political matter. We have to co-exist. Islam is closer to Jewry than Christianity. If Jews and Christians can live together, there's no reason why Jews and Muslims can't. Israel is not going to go away.’ (Times: BoE text NB1-010907)

(2)

By the same token, Muslims cannot be anti-semites. The early Muslims - the Arabs - were themselves Semites. (Independent: BoE text NB2-990113)

There was only one instance of semitism, referring to a subject of academic study.

The BoE British press yielded 45 instances of semitic alone or in combinations other than anti-semitic. Of the 39 occurrences of semitic alone, the majority (18) referred to semitic languages while one alluded to semitic culture. A further 12 used the term in relation to ethnicity. Interestingly, one of these identifies the Palestinians as “Semitic”:

(3)

This country may be described as anti-semitic because it persecutes not only the Palestinians, who are a Semitic people, but Jews like myself. (Guardian: BoE text NB3-990515)

While one instance of semitic refers explicitly to the Jewish religion, several others clearly include people other than the Jews, as for instance this account, apparently of a school nativity play:

(4)

Jesus, Joseph and Mary are a Semitic family - Joseph can be played as an Arafat lookalike, and Mary must be the Jewish mother of all Jewish mothers. It can thus be a Middle East peace parable. (Independent: BoE text NB2-951228)

A further 8 instances of the adjective are in the context of Semitic Studies or scholarship, where it is not clear whether language or culture, or both, are intended.

The compound forms consist of philo-semitic (2) and pro-semitic (1) which appear to be used as antonyms for anti-semitic. The remaining items are one occurrence each of the rather curious psycho-semitic and ultra-semitic.

The compounds with the prefix anti- yielded somewhat more data, as shown in Table 1:

Table 1. Raw frequencies for anti-semit* [3] compounds in UK data.

anti-semite49
antisemite2
anti-semites24
antisemites0
anti-anti-semites1
anti-semitic355
antisemitic8
anti-semitism376
antisemitism6

Many of these co-occur with Jews or Jewish; the only overtly more inclusive usage is in the instance already noted in (2) above: “Muslims cannot be anti-semites”. We will now examine these collocates more closely by sorting them by t-score.

Since the type semite does not occur at all in the UK data, and because of the splitting of hyphenated words (discussed in 3.1 above), we can be sure that all the collocates of semite in Table 2 are in fact collocates of anti-semite. anti is thus the top collocate:

Table 2. Top 20 collocates of semite in UK press, sorted by t-score.

node word

frequency
of collocate
in corpus

frequency
of collocation
with node word

t-score

anti19533496.993666
an566852274.948545
was1277256142.966857
and3708773192.427689
who48628272.228579
as99219182.032219
racist368441.995819
he122609281.844520
virulent30631.731650
a4003775161.728118
fascist98821.412628
peugeot103521.412552
lifelong141021.411951
shelter214821.410766
austria353721.408537
jewish589721.404749
lie684421.403229
claiming715321.402733
committed983521.398429
reported1675721.387319

The only surprising thing in this table is that Jewish does not appear higher in the list, collocating only twice with (anti-)semite despite having an overall frequency of 5,897. Rather bizarrely, Peugeot is thrown up as having a higher t-score, due to an apparently duplicated book review in the Independent (text NB2--950924) entitled by its writer “Anti-semite in a Peugeot". The spelling antisemite only occurs twice in the UK subcorpus: all of its collocates only appear once in combination with it and the results are consequently idiosyncratic (schizoid, for example, has the highest t-score). These results have not been shown here.

In Table 3, the 3 instances of semites are combined with the 24 instances of anti-semites and the 1 of anti-anti-semites:

Table 3. Top 20 collocates of semites in UK press, sorted by t-score.

node word

frequency
of collocate
in corpus

frequency
of collocation
with node word

t-score

anti19533275.191277
and3708773122.075500
were41009951.998197
within4266731.700101
by93509151.693684
as99219151.660564
eliminationist721.414207
fanatical65121.413617
muslims423721.410328
germans569821.408988
jewish589721.408805
white4653421.371537
the9452462181.352981
many13523921.290184
misappropriated4310.999944
supremacists5710.999926
desecrated6210.999920
sulphurous8610.999888
zionism13010.999831
algerians21710.999719

Predictably, anti is once again the top collocate. The fact that muslims has a higher t-score than Jewish is solely attributable to sentence (2), above, in which muslims appears twice, in both the left and right context of anti-semites. The same text is responsible for zionism appearing once as a collocate of semites: the sentence which immediately follows (2) reads “anti-zionism is a separate matter". It is worth noting that what is presented by the software as zionism as a collocate of semite is in fact a reflection of a negative assertion about their opposites: anti-zionism is being dissociated from anti-semitism.

The spelling antisemites does not occur.

Table 4 gives collocates for semitic, including all hyphenated combinations, while Table 5 gives collocates for the unhyphenated spelling antisemitic:

Table 4. Top 20 collocates of semitic in UK press, sorted by t-score.

node word

frequency
of collocate
in corpus

frequency
of collocation
with node word

t-score

anti1953332618.037680
remarks3657204.458689
racist3684163.984855
and37087731003.901281
an566852313.893603
languages218482.815730
made147217122.765265
nazi308472.626583
literature379872.622146
studies763562.398234
he1226092332.234827
farrakhan32752.233663
jews365452.209196
or358844142.164591
views1054252.158542
violence1291352.141106
were410099152.131772
as992191272.056209
tirade28841.997632
semitic35541.997081
persecution90041.992600
notoriously134641.988933
jewish589741.951515

Table 5. Top 20 collocates of antisemitic in UK press, sorted by t-score.

node word

frequency
of collocate
in corpus

frequency
of collocation
with node word

t-score

he122609241.772824
was127725631.458784
hatred207221.413671
american6514921.397142
have84875621.191812
i105422921.137971
degenerating4210.999984
engulfs5010.999981
protocols15510.999943
ferociously27010.999900
texts83610.999690
coherent104610.999612
wagner123710.999542
admittedly175110.999351
propaganda215410.999202
hated225210.999165
tide250410.999072
expression323910.998800
tragedy631710.997659
exercise850110.996850

In this case it is hard to disentangle the collocates of the 355 instances of anti-semitic from the collocates of the 45 instances of semitic alone or in combinations other than anti-semitic.  The unhyphenated form antisemitic only occurs 8 times and so the data in Table 5 does not add much information.

A review of the concordance listings confirms our intuition that languages and studies are collocates of semitic, while the proper name Farrakhan (a black Muslim leader who was accused of anti-semitism)  appears only in the context of anti(-)semitic, like Wagner (a composer popular with the Nazis). We may note in passing that semitic appears to collocate with itself due to examples like (5):

(5)

Whether a stereotype is anti-Semitic or philo-Semitic is, above all, determined by a wider social context. (Guardian: BoE text NB3-950519)

We can observe that, while Jews and Jewish appear in the most frequent collocates of this ‘hybrid’ type, neither Israel nor Zionism does.

Table 6. Top 20 collocates of semitism in UK press, sorted by t-score.

node word

frequency
of collocate
in corpus

frequency
of collocation
with node word

t-score

anti1953338519.604032
of41786121677.276142
in3131430984.375529
and37087731023.686648
racism3563113.297864
jewish589792.965673
he1226092382.691031
jews365472.621633
poland542272.609964
was1277256382.546090
accused1630172.538158
victim1025562.376379
dreyfus23952.234201
virulent30652.233678
evidence2808352.016747
intolerance48341.995783
913475941.993373
rife79741.993041
nazis158041.986204
holocaust160541.985986

Since there is only one instance of semitism in the BoE English News, almost all the collocates in Table 6 belong to the 376 instances of anti-semitism. (There are also 6 instances of antisemitism which is counted as a separate type). Not surprisingly, therefore, Jewish and Jews are prominent collocates, as are word types indicating notorious historical episodes of persecution of Jews such as Dreyfus, Nazis and holocaust. Here again there is no evidence of lexis relating to Israel or Zionism.

The US press was less forthcoming with instances of these items. There were no occurrences whatsoever of semite, semites, semitism or semitic. This may, of course, be attributable to the relatively small size of our US sub-corpus (10m as opposed to 173m tokens).There was one occurrence of anti-semite (none of the plural or unhyphenated variants):

(6)

In the 1992 campaign, he noted that he had been called ‘an anti-Semite, a homophobe, a racist, a sexist, a nativist, a protectionist, an isolationist and a beer-hall conservative.’ Not true, he said: ‘I am none of the above.’ (USnews: BoE text NU1-960223)

The remaining instances of compound anti- forms comprised 6 occurrences of anti-Semitic, 1 of antisemitic and 14 of anti-semitism. It is often clear from the contexts that the terms relate to the treatment of Jewish people; however, it is also clear that the use of these terms is hotly contested, as in the following example which uses both antisemitic and anti-Semitism in the same passage:

(7)

The Post yesterday noted that Buchanan is the only major party presidential candidate who has been the subject of reports by the Anti-Defamation League, which has said his politics are ‘defined by prejudice and rancor, if not outright hate.’ Buchanan has called the ADL's reports about him an ‘orchestrated smear campaign.’ ‘If these comments stood alone, we might be reading too much into them,’ said Abraham Foxman, national director of the ADL. ‘Unfortunately, this comes with antisemitic, anti-Israel, Holocaust-denial baggage and exclusionary, extremist language. He is a lot more careful in word selection recently, but to this day he has never repudiated his language nor apologized for it.’ However, The Post also quoted Buchanan's former debating partner on CNN's ‘Crossfire,’ Michael Kinsley - one of several Buchanan colleagues who have defended him against anti-Semitism allegations. ‘As a Jew,’ Kinsley has said, ‘I never felt any hostility from Buchanan on that score, never heard him make a disparaging remark about Jews, never noticed any difference in the way he treats Jews and non-Jews.’ (USnews: BoE text NU1-960223)

The quotation attributed to Foxman juxtaposes the terms antisemitic, anti-Israel and Holocaust-denial, arguably implying a degree of equivalence, while the words attributed to Kinsley appear to restrict the meaning of anti-Semitism to hostility towards, or unequal treatment of, Jewish people.

We have omitted t-score data derived from single occurrences of word types (anti-semite and antisemitic) in the US data, since such tables add nothing to the concordance information.

Table 7. Top 20 collocates of semitic in US press, sorted by t-score.

node word

frequency
of collocate
in corpus

frequency
of collocation
with node word

t-score

anti127172.643446
was4076621.275885
with5239421.236429
gritz110.999995
thicket710.999966
rubbing1210.999942
bo2810.999866
racist5810.999722
replied12110.999419
literature14510.999304
jews14810.999290
feared16710.999199
remarks20510.999016
primary79410.996190
charged81410.996094
played94610.995460
rules118410.994318
example175210.991593
asked193810.990700
white280610.986535

Table 8. Top 20 collocates of semitism in US press, sorted by t-score.

node word

frequency
of collocate
in corpus

frequency
of collocation
with node word

t-score

anti1271143.737854
allegations37131.729652
amounting2921.413984
racism7721.413604
reports139221.403192
growing151021.402258
things231121.395916
his2244521.236505
he3955721.101020
from4292821.074330
of24067351.030903
gentry710.999922
impatience1110.999877
sprang1410.999843
novak1810.999798
spectator2010.999776
jew3110.999653
dismisses3610.999597
manchester4710.999474
nonsense5410.999395

As with the UK data, anti appears artificially as a collocate of semitic and semitism in Tables 7 and 8. There is no sign of any reference to Zionism or Israel in these data, only to items such as Jew and racist.

3.1.2 Zionism, Zionist, anti-Zionism and anti-Zionist

There were copious examples of variants of Zion with various prefixes and suffixes in the UK Press (there were no unhyphenated forms of any of the compound words; capitalized and non-capitalised spellings have been grouped together):

Table 9. Raw frequencies for Zion* compounds in UK data.

Zionism122
anti-Zionism8
Zionist183
Zionists57
pro-Zionist1
non-Zionist5
post-Zionist1
anti-Zionist15
anti-Zionists2

While the terms Zionist, anti-zionist and their variants do not seem to be disputed terms in themselves, those employing them often seem to be taking pains to contrast them with other terms such as anti-Judaic or anti-Semitic, as in example (8):

(8)

Prejudice can be religious, ie anti-Judaic; it can be racist, ie anti-Semitic; and it can be political, ie anti-Zionist. Prejudice may combine all three, but one prejudice does not automatically assume the other two. There may be those who oppose the political ideas of Zionism, but are not either anti-Judaic or anti-Semitic. (Guardian: BoE text NB3-990116)

The US data is once again somewhat sparse, yielding only the following forms (as with the UK data, there were no unhyphenated forms of the compounds):

Table 10. Raw frequencies for Zion* compounds in US data.

Zionism1
Zionist4
Zionists5
anti-Zionist2

The term Zionist is not always used to denote individuals: one of these examples quotes a spokesman for the military wing of Hamas referring to Israel as “the Zionist entity.” (BoE text NU1-960305)

While one of the two instances of anti-Zionist refers to Algerians as “victims of their own country's anti-Zionist propaganda” (BoE text NU2-890906), the other somewhat surprisingly turns out on closer examination to be concerned with haredi (ultra-orthodox) Jews:

(9)

Twenty years ago, the haredi were almost completely uninvolved in secular politics, because of their belief that no Jewish state should exist until after the Messiah comes. But 10 years later, even many of the anti-Zionist sects had accepted politics as something of a necessary evil – although they would vote only for ultraorthodox political parties and limited their involvement almost exclusively to religious issues such as banning public transportation on the Sabbath, ensuring that only orthodox marriages are recognized by the state, and securing funding for ultraorthodox institutions. (USnews: BoE text NU1-960526)

Tables 11–13 show the top collocates of Zionism, Zionist and Zionists respectively in the UK press sub-corpora.

Table 11. Top 20 collocates of Zionism in UK press, sorted by t-score.

node word

frequency
of collocate
in corpus

frequency
of collocation
with node word

t-score

and3708773534.212390
racism3563153.867444
of4178612533.823762
anti19533123.430147
israel10551113.297468
that1547837232.852331
was1277256192.594387
jewish589752.220187
equating6341.999810
resolution302941.990880
political4982141.849995
state5409241.837136
international5884741.822819
<h>6486641.804696
</h>6496841.804389
is1726001181.792855
for1615098171.764269
judaism27531.731095
imperialism35831.730806
islam270831.722636

Table 12. Top 20 collocates of Zionist in UK press, sorted by t-score.

node word

frequency
of collocate
in corpus

frequency
of collocation
with node word

t-score

the94524621535.112718
anti19533204.430661
of4178612774.253069
israel10551113.286416
a4003775633.147279
entity76092.997594
zionist20582.827739
state5409282.646825
non2412272.559175
which419924132.499608
conspiracy289262.438278
american6514972.411925
israeli674052.207445
regime766852.203504
religious780752.202914
movement1174852.186178
party8289162.128149
aggression185341.991202
lobby311841.985196
federation469741.977699

Table 13. Top 20 collocates of Zionists in UK press, sorted by t-score.

node word

frequency
of collocate
in corpus

frequency
of collocation
with node word

t-score

the9452462543.833026
jews365452.231602
were41009972.222136
for1615098122.189894
by93509192.148147
most18692241.744576
not64915861.725209
religious780731.719732
anti1953331.701230
<p>148231091.649639
who48628251.641728
men6601931.627881
their51756251.603497
against14920831.496620
being15359631.489696
other18364731.442280
zionists5921.414100
reciprocated12121.413980
despised40921.413423
orthodox215021.410059

While anti appears as a collocate of all three words, reflecting the presence of forms like anti-zionism, there is no evidence of anti-semitism or its related forms being a significant collocate of (anti)zionism. (We noted that terms such as antisemitism did occur in some of the texts thrown up by the software, but not usually within a span of ±4 relative to the node word, so that these did not appear among the top 20 collocates).

However, racism does appears as a significant collocate, co-occurring with zionism 15 times within the 4-word span. It occurred to us that racism could be being used as a synonym or superordinate of anti-semitism in these contexts, and so we returned to the concordance listings to investigate this possibility. In fact we found that without exception the collocation of racism with zionism was in contexts such as (10):

(10)

On September 23rd Mr Bush asked the UN General Assembly to repeal the resolution it adopted in 1975 describing Zionism as a form of racism. (Economist: BoE text MB1-910928)

Thus it is Zionism itself, and not opposition to it, which is being equated with racism in these texts. The equation is a highly contested one, of course, with the UN resolution in question being labelled a “vile canard” and a “libel” by some of the speakers quoted.

Due to the smaller size of the US sub-corpus the collocate data here are somewhat sparse. Since Zionism only occurs once in our US data we have omitted the t-score data for this type. Tables 14 and 15 give the collocates for Zionist (including anti-Zionist) and Zionists respectively.

Table 14. Top 20 collocates of Zionist in US press, sorted by t-score.

node word

frequency
of collocate
in corpus

frequency
of collocation
with node word

t-score

accepted57721.412256
anti127121.409901
he3955721.279988
samuil210.999990
leaflet410.999981
sects510.999976
propaganda6610.999683
mighty10210.999511
entity13310.999362
referred17210.999175
writes19310.999074
conspiracy21410.998973
salaries22710.998911
rarely23510.998872
attacks28110.998652
inside82210.996055
indeed83410.995998
politics86210.995863
created98010.995297
perhaps110210.994712

Table 15. Top 20 collocates of Zionists in US press, sorted by t-score.

node word

frequency
of collocate
in corpus

frequency
of collocation
with node word

t-score

the50483151.333235
for10056021.129861
flitted110.999996
andreevna310.999988
italians2810.999888
hungarians3010.999880
turks3210.999872
vera5210.999792
crush6710.999732
blames7210.999712
russians10110.999596
swing24810.999008
germans25310.998988
recognize25410.998984
israel58410.997665
love93810.996249
wouldn155110.993798
ms163310.993470
campaign183310.992670
nation190510.992382

The low frequencies of both the node words and the collocates in the US data may well mean that the patterns shown here are idiosyncratic. Nonetheless we note that there is no evidence of “anti-semitism” or its variants featuring as a prominent collocate of “anti-zionism” or its variants.

3.1.3 Summary of BoE findings

Our survey of BoE news discourse does indicate some possible differences between US and British usage of the items in question. There is evidence that semites and semitic are in use in current British English to denote Arabs as well as Jews, and indeed to emphasize the common ethnic heritage of the two groups. We found no evidence of these items at all in our US sample, which may simply be due to its much smaller size (10m word tokens, as opposed to 173m in its UK counterpart).

In both the UK and the US data, the terms anti-semite, anti-semitic and anti-semitism (in various spellings) are used almost exclusively in relation to Jews. It could be stated that the lexical items incorporating the negative prefix anti- are more restricted in their semantic domain. Nonetheless they are the site of ideological struggle, with both British and American speakers attempting to reject the application of such terms to particular individuals, and British speakers in particular taking pains to distinguish anti-semitism from anti-Zionism. Zionism and its compounds do not seem to be frequent items in US news discourse, even if the size of the corpus is taken into account.

Our analysis of the BoE yielded no evidence that the meaning of antisemitism in recent media discourse has moved towards including anti-Zionism. However, the BoE does not provide the facility to observe shifts in usage over time. In order to attempt this we needed to turn to a different source of data.

3.2 WebCorp

3.2.1 Overview

Next, we decided to focus specifically on the words anti-Semitism, anti-Semitic and anti-zionism and to trace their evolution over the past few years by means of a diachronic collocational study of news texts drawn from the World-Wide Web.

The tool we chose for this part of the study was WebCorp's Linguistic Search Engine, as developed at Birmingham City University (Renouf et al. 2007). WebCorp is a software suite which in effect turns the World-Wide Web into a huge corpus. It ingeniously ‘piggy-backs’ onto a standard search engine such as Google and extracts data from web pages, allowing the user to specify search terms.

We utilised all the newspaper sources that can be accessed systematically through WebCorp. For the UK this comprised The Times, The Daily Telegraph, The Guardian, The Independent, The Mirror, The Sun, The News of the World and The Daily Star. For the US, the available newspapers were The New York Times, The Washington Post, The Los Angeles Times and Suntimes [4]. We shall henceforth refer to these sources as “UK Press” and “US Press” respectively; but it should be noted that they do not exactly match the categories of the same names in the BoE.

Since both anti-semitism and anti-zionism are subject to various spelling permutations we used the search patterns:

[antisemitism|anti-semitism]

[antizionism|anti-zionism]

In combination with the option ‘Case-insensitive’ these templates cover all possible spellings (with or without capitals, with or without hyphens etc.).

Our first trawl through the ‘UK Press’ and ‘US Press’ took no account of the dates when the web pages concerned originated: we searched all that was accessible in the respective domains, but were restricted by WebCorp to a maximum of 500 pages of input. In all searches we used Google as the default search engine.

Tables 16–19 show the collocational grids derived from the concordances for anti-semitism and anti-zionism in the UK and US sources respectively, showing the 20 most frequent collocates [5] and their position relative to the keyword (with a span of up to 4 words to the left or right) in descending order of frequency. WebCorp does not currently provide tools for calculating MI or t-score statistics, and so all figures in this section are raw frequencies.

Table 16. Collocates of antisemitism in the UK press (raw frequencies), non-date-specific.

Word

Total

L4

L3

L2

L1

 

R1

R2

R3

R4

Left Total

Right Total

Israel

116

12

14

64

1

2

4

7

12

91

25

Jews

101

13

17

13

1

7

13

18

19

44

57

article

101

13

3

12

47

18

8

28

73

new

98

10

16

2

64

3

2

1

92

6

Europe

83

8

8

1

7

46

6

7

24

59

rise

71

3

5

39

3

3

18

47

24

said

69

10

4

6

5

8

16

13

7

25

44

Buzz

68

30

21

9

8

0

68

Jewish

64

12

3

4

2

11

13

19

19

45

anti-Zionism

61

1

8

30

15

4

3

39

22

first

59

6

1

4

1

47

7

52

racism

55

2

5

23

6

4

12

2

1

36

19

Digg

52

1

30

21

1

51

rising

46

1

16

15

3

8

3

32

14

France

45

3

1

3

26

6

6

4

41

accused

42

2

11

27

1

1

40

2

European

38

3

5

21

2

1

2

4

29

9

UK

37

1

3

2

9

4

18

15

22

Islamic

36

5

1

25

2

3

31

5

Close

36

13

10

2

11

36

0

Table 17. Collocates of antisemitism in the US press (raw frequencies), non-date-specific.

Word

Total

L4

L3

L2

L1

 

R1

R2

R3

R4

Left Total

Right Total

Published

147

72

14

5

56

0

147

Autos

130

49

43

9

29

130

0

E-MAIL

99

62

5

21

11

0

99

Print

85

2

6

55

6

16

8

77

Estate

81

43

9

29

81

0

Jews

74

17

12

15

6

2

5

9

8

50

24

Reprints

69

4

2

6

30

27

12

57

said

65

10

6

16

7

8

11

7

32

33

new

57

3

26

1

20

3

4

50

7

New

51

3

22

2

4

1

19

25

26

France

48

5

7

12

3

20

1

24

24

racism

47

1

3

11

19

13

15

32

Israel

45

7

5

20

2

11

32

13

rise

45

11

22

7

5

40

5

Muslim

42

1

26

1

4

10

27

15

Poland

39

6

5

3

1

23

1

15

24

Real

38

9

29

38

0

world

38

16

16

6

16

22

Jewish

37

6

4

5

4

2

4

8

4

19

18

Europe

36

6

13

6

7

4

25

11

Table 18. Collocates of antiZionism in the UK press (raw frequencies), non-date-specific.

Word

Total

L4

L3

L2

L1

 

R1

R2

R3

R4

Left Total

Right Total

anti-semitism

97

4

32

2

2

43

14

38

59

anti-Semitism

40

4

3

14

2

12

3

2

23

17

antisemitism

25

3

1

19

2

4

21

Jewish

19

5

6

3

5

11

8

Israel

17

1

5

2

2

7

6

11

become

15

1

4

10

0

15

anti-semitic

14

1

2

2

3

4

2

8

6

Jews

14

3

4

2

3

2

9

5

conflating

12

12

12

0

Stop

12

12

12

0

criticism

11

4

4

3

4

7

difference

11

9

2

11

0

distinction

10

8

2

10

0

Distinguishing

9

9

9

0

now

8

1

1

5

1

7

1

Editorial

8

8

8

0

masquerading

8

8

8

0

2009

8

4

1

3

8

0

confuse

8

8

8

0

anti-Semitic

8

7

1

0

8

Table 19. Collocates of antiZionism in the US press (raw frequencies), non-date-specific.

Word

Total

L4

L3

L1

L1

 

R1

R2

R3

R4

Left Total

Right Total

anti-Semitism

71

6

1

19

1

31

11

2

26

45

Israel

27

8

1

9

2

5

1

1

18

9

hate

17

17

0

17

anti-semitism

16

6

1

3

3

3

6

10

Jewish

15

1

5

4

1

4

11

4

Opinion

14

8

6

8

6

rejects

12

8

4

0

12

dangerous

12

8

4

12

0

character

12

4

4

4

8

4

distinction

12

5

7

12

0

say

10

7

1

1

1

9

1

spoke

9

9

9

0

last

9

1

8

0

9

link

8

8

8

0

become

8

6

2

0

8

said

8

1

4

3

1

7

Palestinian

8

1

4

3

8

0

discriminatory

8

4

4

0

8

conference

8

8

0

8

Tehran

8

8

0

8

Unfortunately Tables 16 and 17 include a number of ‘artefacts’, presumably resulting from the way the web pages are accessed and processed by WebCorp. In the US Press antisemitism grid the top 5 items and the 7th (Published/Autos/E-MAIL/Print/Estate and Reprints) can be discarded as ‘noise’ of this kind. In the UK Press, Buzz and Digg and probably article are the intruders. Ignoring these items, we can see that some words which occur in the UK results match those in the US results, but occur in different relative positions. Most strikingly Israel is at the top of the UK grid, but considerably lower down in the US one. The word rise is higher in the UK than the US data, while rising only occurs in the UK grid. Not surprisingly, Europe is higher in the UK chart. By contrast, Jews and Jewish rank roughly equally in both, and the same applies to racism and new/New, which will be discussed in some detail later. Finally we may observe that the UK grid contains some words that are not found in the US one, most notably anti-Zionism, but also accused and Islamic (where US has Muslim).

Turning to the grids for antizionism, there seems to be a somewhat closer correspondence at the top between the UK and US results than for antisemitism. Most strikingly, in both the UK and the US grids for antizionism, anti-semitism in its various permutations ranks highest (though more so in the UK, which also includes anti-semitic and Anti-semitic, which are absent from the US data). Israel is high in both, and so is Jewish (plus Jews in the UK data). Distinction also occurs in both with roughly equal weight, if we also include Distinguishing, which only occurs in the UK Press. For the rest there are more differences than similarities.  Notice in particular hate, high up in the US chart, and Palestinian and discriminatory towards the bottom end, compared with such items as conflating, confuse and masquerading in the UK grid.

It may be of some interest to note that, judging from the above four charts, antizionism seems to ‘attract’ anti-semitism, but the converse does not apply. In other words, there seem to be far more media web pages, and/or sentences within them, in which anti-semitism occurs without having anti-zionism in its immediate context, than vice-versa. When one examines the actual concordance lines it is noticeable that there is a lot of debate about whether antizionism is in fact always anti-Semitic. Claims that antizionism is anti-Semitism are countered by counterclaims that antizionism is not anti-Semitism: witness the following sample of concordance lines extracted from the US Press using the pattern ‘[antizionism|anti-zionism] is not’:

(11)

http://www.latimes.com/la-oe-pearl15-2009mar15-story.html
The organizers, some of them Jewish, took refuge in ‘academic freedom’ and the argument that anti-Zionism is not anti-Semitism. we fully support this mantra, not because it exonerates anti-Zionists from charges of anti-Semitism but because the distinction helps us focus attention on the discriminatory …

(12)

http://newsweek.washingtonpost.com/onfaith/panelists/adin_steinsaltz/2007/02/criticizing_israel_without_bei.html[site no longer available, see https://web.archive.org/]
Are you deliberately trying to mislead people into equating anti-semitism with anti-zionism? No, anti-Zionism is not anti-semitism.  As an idea, a Jewish homeland was always controversial. As a reality, Israel still is - and it is not anti-Jewish to say so …

(13)

http://newsweek.washingtonpost.com/onfaith/2007/02/criticism_of_israel/comments.html
[site no longer available, see https://web.archive.org/]
IT IS POSSIBLE TO QUESTION A GOVERNMENT WITHOUT HOLDING THE PEOPLE LIVING IN IT RESPONSIBLE FOR ITS ACTIONS- ANTI-ZIONISM IS NOT EQUAL TO ANTI-SEMITISM ANTI-SEMITISM IS EQUAL TO ISLAMOPHOBIA (the benign and pretty sounding name for anti-muslim) even the name downplays the hate factor hate is hate and racism is racism …

Our next step was to produce collocates for the adjectival forms antisemitic and anti-Semitic, since these were expected to exhibit different collocational patterns from their nominal equivalents. Tables 20 and 21 give the ‘full trawls’ for the UK and US press respectively.

Table 20. Collocates of antisemitic in the UK press (raw frequencies), non-date-specific.

Word

Total

L4

L3

L2

L1

 

R1

R2

R3

R4

Left Total

Right Total

attacks

122

116

2

4

0

122

incidents

118

1

1

111

5

2

116

racist

65

6

26

6

5

17

5

38

27

Jewish

65

13

12

15

3

4

12

6

40

25

Israel

62

11

15

18

3

3

2

8

2

47

15

article

61

10

3

8

2

3

16

11

8

23

38

play

53

17

10

11

4

7

1

1

2

42

11

said

49

8

4

2

5

1

10

12

7

19

30

remarks

48

1

4

1

41

1

6

42

cartoon

46

5

1

10

14

3

3

10

30

16

number

45

2

3

37

3

42

3

Europe

40

4

2

1

22

11

7

33

France

39

7

3

5

18

6

15

24

rise

37

3

28

1

5

32

5

Britain

31

3

2

2

24

7

24

Jews

30

6

1

4

1

8

10

11

19

propaganda

30

7

1

20

2

8

22

Buzz

30

7

9

12

2

0

30

accused

27

9

14

4

27

0

attack

27

2

1

1

4

17

1

1

8

19

Table 21. Collocates of antisemitic in the US press (raw frequencies), non-date-specific.

Word

Total

L4

L3

L2

L1

 

R1

R2

R3

R4

Left Total

Right Total

said

103

17

9

4

1

2

21

25

24

31

72

Autos

88

20

31

2

35

88

0

incidents

78

2

3

5

67

1

10

68

Estate

68

31

2

35

68

0

remarks

63

2

2

57

1

1

4

59

made

62

21

26

13

1

1

47

15

comments

55

1

4

48

1

1

5

50

E-MAIL

54

3

30

11

10

0

54

remark

49

2

2

45

4

45

acts

46

3

43

3

43

Jewish

42

5

3

2

1

7

16

8

10

32

Published

41

10

1

12

7

11

10

31

Print

41

2

1

8

2

18

10

11

30

racist

39

3

1

25

2

7

1

29

10

France

37

1

2

5

10

17

2

18

19

graffiti

37

1

35

1

1

36

Real

37

2

35

37

0

Jobs

35

35

35

0

Reprints

32

1

2

1

8

2

18

12

20

Israel

32

11

6

6

2

2

4

1

25

7

Ignoring the same ‘junk’ words, here again we see that Israel is high up in the UK chart, but lower down for the US, while Jewish is high up in both. Three further shared and fairly similarly ranked items are incidents, remarks and racist. Notice finally that attacks, almost always occurring one position to the right of the node, tops the UK chart, while being completely absent in the US data, where the closest in meaning is the more neutral term acts.

3.2.2 Impact of external events on collocational patterns

Since we expect developments in the ‘real world’ to be reflected in discourse practices, we segmented the concordances into yearly ‘slices’ for 2006 and 2007 (the only years which generated sufficient data for this to be viable). We were interested to see whether the distribution of collocates would be affected, for instance, by the appearance of the EUMC ‘working definition of antisemitism’ in 2005 and, in the UK, the All-Party Parliamentary Inquiry into Antisemitism, which delivered its report in 2006. Tables 22 and 23 show the UK and US grids respectively for 2006 while Tables 24 and 25 show the corresponding grids for 2007:

Table 22. Collocates of antisemitism in the UK press (raw frequencies), 2006.

Word

Total

L4

L3

L2

L1

 

R1

R2

R3

R4

Left Total

Right Total

back

9

1

8

0

9

Israel

6

2

1

1

2

2

4

wave

6

6

6

0

form

5

3

2

3

2

September

5

5

5

0

definition

5

1

1

3

5

0

report

5

1

3

1

5

0

just

5

1

2

1

1

1

4

European

4

2

2

2

2

increase

4

2

1

1

3

1

Jewish

4

1

1

2

1

3

see

3

1

2

3

0

positive

3

1

1

1

1

2

manifestation

3

2

1

3

0

global

3

2

1

2

1

Denis

3

2

1

2

1

people

3

2

1

3

0

new

3

2

1

3

0

GBR

3

3

3

0

growth

3

1

2

1

2

In offering our “interpretation” (Fairclough 2001: 21) of the processes whereby such distributions are produced and interpreted, we maintain that the prominence of Israel as a collocate of antisemitism cannot be dissociated from the appearance of other collocates such as definition, report, European and Denis (undoubtedly Denis MacShane who chaired the all-party parliamentary inquiry). The ‘pole position’ of the word back is also explained indirectly by the APPG enquiry report: its launch in September 2006 was heralded by an article from MacShane on the Guardian's Comment is Free website, rather provocatively entitled ‘Anti-semitism is back’. The report itself controversially recommended the adoption of the EUMC “working definition of antisemitism” of the previous year, which included “Denying the Jewish people their right to self-determination, e.g., by claiming that the existence of a State of Israel is a racist endeavour” (APPG 2006: 6). There were numerous other references to Israel both in the EUMC definition and in the report, and the press interest in quoting from both of these seems to have impacted on the relative frequencies in our collocation charts.

None of this is mirrored in the US Press for the same year, whose discourse concerning antisemitism is focussed on the Holocaust, reflected in the use of collocates such as Poland, Polish, Auschwitz, World, War, terrible and fear.

Table 23. Collocates of antisemitism in the US press (raw frequencies), 2006.

Word

Total

L4

L3

L2

L1

 

R1

R2

R3

R4

Left Total

Right Total

Jews

11

2

3

3

3

8

3

Poland

11

3

8

3

8

Auschwitz

8

8

0

8

Fear

6

6

6

0

paragraph

5

2

2

1

4

1

War

5

2

3

0

5

World

5

2

3

0

5

today

4

4

4

0

terrible

4

1

3

1

3

Catholic

4

3

1

3

1

2004

4

1

2

1

4

0

Related

4

1

3

4

0

Roman

4

3

1

3

1

Searches

4

1

3

4

0

Articles

4

1

3

1

3

Literature

3

1

2

3

0

Polish

3

3

3

0

Books

3

2

1

2

1

next

3

2

1

2

1

cured

3

3

3

0

By 2007 the focus in the UK Press had shifted to academia, with collocates such as education, students, campus and universities. One's first impression is that there must have been a dramatic rise in attacks on Jewish students at British institutions that year. However, the occurrence of all-party as a prominent collocate suggests an alternative interpretation: the report of the all-party inquiry had devoted a whole chapter to “Antisemitism on Campus” (APPG 2006: 38-42), as well as a sub-section on “anti-Zionism” under “Antisemitic discourse”. Anti-Zionism, it will be seen, is another collocate which features here.

Table 24. Collocates of antisemitism in the UK press (raw frequencies), 2007.

Word

Total

L4

L3

L2

L1

 

R1

R2

R3

R4

Left Total

Right Total

difference

3

1

1

1

2

1

education

3

1

1

1

2

1

all-party

2

1

1

2

0

guilty

2

1

1

1

1

said

2

1

1

0

2

Students

2

1

1

2

0

campus

2

2

2

0

anti-Zionism

2

1

1

1

1

Letters

2

1

1

2

0

report

2

1

1

2

0

Higher

2

1

1

1

1

universities

2

1

1

0

2

thing

2

1

1

0

2

grain

1

1

1

0

provoke

1

1

1

0

rise

1

1

1

0

Opinion

1

1

1

0

another

1

1

1

0

form

1

1

1

0

take

1

1

1

0

As the all-party inquiry was a purely British phenomenon, we would not expect its discourse to be reflected in the US data, and indeed it is not: the top collocate in 2007 is Jews as in 2006, while the emphasis on the Holocaust is no longer apparent. This is probably because the year 2005 marked the 60th anniversary of the liberation of Auschwitz and the events commemorating it would still have been in the public eye the following year, while by 2007 they would have become less prominent in the media. It is striking, however, that these events (which received plenty of attention in the British media) have not been reflected in our collocation charts for the UK press in 2006, the lexis of the Shoah having apparently been edged out by that of a parliamentary inquiry. An “explanation” of such phenomena in Fairclough's sense (2001: 21) has to take into account the wider social processes at work here.

Table 25. Collocates of antisemitism in the US press (raw frequencies), 2007.

Word

Total

L4

L3

L2

L1

 

R1

R2

R3

R4

Left Total

Right Total

Jews

11

3

3

2

3

8

3

New

11

8

3

8

3

Thought

8

8

8

0

rise

6

6

6

0

world

6

3

3

3

3

praised

3

3

0

3

detractors

3

3

0

3

historian

3

3

0

3

vicious

3

3

3

0

subject

3

3

3

0

contributing

3

3

3

0

virulent

3

3

3

0

advocacy

3

3

0

3

number

3

3

0

3

feeding

3

3

3

0

questioning

3

3

0

3

group

3

3

0

3

references

3

3

3

0

says

3

3

0

3

said

3

3

0

3

One collocate which makes a sudden entry in the US chart in 2007, and goes straight to the top (taking joint first place with Jews), is the word new, predominantly in the position one slot to the left of the node word, i.e. forming the phrase new antisemitism. Some of the occurrences accounting for this pattern are as follows:

(14)

https://www.onfaith.co/onfaith/2007/02/21/israel-is-a-nationstate/834
…read in that newspaper. This is what many call the “new anti-Semitism” of the left, but it is also found on the …

(15)

https://www.onfaith.co/onfaith/2007/02/21/israel-is-a-nationstate/834
… of the very idea of a Jewish state. But this ‘new anti-Semitism’ draws on the old stereotypes of ‘Jewish conspiracies’ and the …

(16)

http://newsweek.washingtonpost.com/onfaith/2007/02/criticism_of_israel/comments.html[site no longer available, see https://web.archive.org/]
… to go beyond mere assertion, claim that there is a ‘new anti-semitism’, which they equate with criticism of Israel. In other words, …

The phrase the new antisemitism became prominent in 1974 when two officials of the Anti-Defamation League published a book of that title (Forster & Epstein 1974). US scholar Norman Finkelstein gives short shrift to the concept of ‘the new anti-Semitism’, alleging that it consists of three components: (i) “exaggeration and fabrication", (ii) “mislabeling legitimate criticism of Israeli policy,” and (iii) “the unjustified yet predictable spillover from criticism of Israel to Jews generally” (Finkelstein 2005: 66). According to Finkelstein, claims of a ‘New Antisemitism’ arise about every 15 years and are in fact nothing new at all. They seem to have reappeared in the US press in 2007, at any rate, but the phrase was not in sufficient circulation on the other side of the Atlantic to have any impact on the UK collocation chart for that year. WebCorp, then, has flagged up a further instance of contested language in this fraught domain of discourse.

Finally we wished to explore further the impact of the EUMC Report on Antisemitism published in March 2004, and the “working definition of anti-Semitism” arising from that Report which the EUMC adopted the following year. For this we used the pattern:

[antisemitism|anti-semitism] with additional filter: [EUMC|EUMC's]

This returned ‘hits’ when the word EUMC or EUMC's appeared anywhere in the same web page as antisemitism. The two words were not necessarily collocates: in some cases they had even been used by different authors, such as when someone had written a ‘blog’ and readers had posted comments in response to it, as is the case with Denis MacShane's piece in The Guardian's “Comment is Free” blog, examples 2023 below.

This pattern generated 354 concordance lines in the UK Press, including the following examples, not all of which immediately strike the reader as being connected with the EUMC:

(17)

https://www.theguardian.com/education/2007/mar/29/highereducation.uk1
EUMC, now the European Union Agency for Fundamental Rights, defines anti-semitism as the expression of hatred towards Jews, their property and …

(18)

http://www.guardian.co.uk/education/2007/mar/29/highereducation.uk1
… this involves combating all forms of intolerance on campus, including anti-semitism’. He added: “Universities have a legal obligation to ensure academic …

(19)

https://www.theguardian.com/world/2004/apr/01/religion.race
… to be young, disaffected white Europeans. ‘A further source of anti-semitism was young Muslims of North African or Asian extraction.’ Traditionally …

(20)

https://www.theguardian.com/commentisfree/2006/oct/12/post503
… on p14 and says ‘Examples of the ways in which anti-semitism manifests itself with regard to the State of Israel taking …

(21)

https://www.theguardian.com/commentisfree/2006/oct/12/post503
… Palestinians. Is that criticism of Israel to be regarded as anti-semitism and become punishable by law ?' You seem keen on free …

(22)

https://www.theguardian.com/commentisfree/2006/oct/12/post503
… that corrupted French political life during the war. The deep anti-semitism of the French which sent many jews to the gas …

(23)

https://www.theguardian.com/commentisfree/2006/oct/12/post503
… aren't you agitating right now for a legal definition of anti-semitism that would encompass criticism of israel such as comparing it …

The collocation grid for that query is shown in Table 26. It is not surprising that EUMC appears as a collocate, given that this was supplied as part of the search context in the first place. What is much more striking is that the word Israel tops the chart and collocates with antisemitism even more strongly than Jews does.

Table 26. Collocates of antisemitism in the context of EUMC in the UK press (raw frequencies), non-date-specific.

Word

Total

L4

L3

L2

L1

 

R1

R2

R3

R4

Left Total

Right Total

Israel

16

2

1

6

1

3

3

9

7

back

15

2

12

1

0

15

Jews

14

3

4

4

3

7

7

wave

14

14

14

0

definition

14

1

1

11

1

14

0

report

13

1

7

1

1

2

1

9

4

rise

11

2

8

1

10

1

present

10

10

10

0

Britain

8

2

5

1

2

6

EUMC

8

3

3

1

1

6

2

criticism

8

2

1

5

2

6

September

7

7

7

0

increase

7

2

2

1

2

4

3

forum

7

1

2

4

1

6

just

7

1

3

1

2

1

6

Polish

7

5

1

1

5

2

new

7

2

2

3

7

0

people

6

2

2

2

6

0

anti-Zionism

6

1

1

2

2

2

4

European

6

1

2

3

3

3

4. Conclusion

The low frequencies of some of the items discussed here, even in corpora of millions of words, indicate that due caution should be exercised in drawing firm conclusions. All the same we believe that the findings presented above support Fairclough's observation (2001: 19) that “Politics partly consists in the disputes and struggles which occur in language and over language”. There is ample evidence here for “disputes and struggles” over the definitions of terms such as semites and anti-semitism.

The disputes sometimes take unexpected forms. Who would have guessed, given the high-profile allegations of Mearsheimer and Walt (2006) about the power of the “Israel lobby” to influence US politics, that the word Israel would hardly feature at all in the collocations of antisemitism in our US data but would repeatedly appear near the top of the charts in our British web corpus? This striking difference would seem to be attributable to the disproportionate influence of the EUMC's Working Definition and its subsequent adoption by the All-Party Parliamentary Group. A number of pressure groups have campaigned – with some success –for this particular definition, which blurs the distinction between anti-zionism and anti-semitism, to be widely accepted and implemented in the EU and beyond, with potentially far-reaching social and even legal implications.

Given the amount of linguistic lobbying which has become apparent from our analysis of the World-Wide Web, lexicographers would be well advised to exercise extreme caution before relying on corpus data in formulating their definitions of contentious words. Organisations and individuals with the resources and know-how to propagate the definitions they favour may actually succeed in skewing the statistics picked up by software such as WebCorp.

We would also urge caution in the use of collocation data for lexicographical purposes. Not only can collocational patterns be distorted just as raw frequencies can be, as we have seen, but there is a risk of overlooking crucial factors such as polarity by relying on ‘stop-word lists’ to exclude function words. It would be all too easy to find an impressive level of collocation between, for example, anti-zionism and anti-semitism, while failing to notice that a large number of the co-occurrences involve the word not.

Webster's New International Dictionary and the EUMC comprise part of the social processes which help to construct contemporary ideology; or as Fairclough puts it, “common sense in the service of power” (2001: 64). Not only public figures like Abe Foxman or Robert Fisk, but the millions of ordinary people whose speech and writing have gone into the corpora we have analysed here, have a part to play in either perpetuating or challenging the linguistic “common sense” which dictionary definitions both shape and reflect.

Notes

[1] The term antisemitism is commonly used both with and without hyphenation, and with and without capitalisation. We will use the spelling anti-semitism except when naming an organisation, quoting a publication or citing a text which uses a different spelling.

[2] The authors would like to thank Tony Greenstein for his assistance with section 2.2.2 on the history of Zionism.

[3] The * symbol is used as a wild card to represent any number of alphanumeric characters. The use of it in a table is merely for brevity and does not imply that it was used in the original search string.

[4] Although WebCorp also offers the option of BBC News as a separate domain, we decided against including this, as there is no corresponding option for American radio or TV sources.

[5] These are termed ‘external collocates’ by WebCorp; the term ‘internal collocates’ is applied to words which occur between two specified node words. We did not conduct any analysis of internal collocates as we were working with a single node word at a time.

References

All-Party Parliamentary Group against Antisemitism. September 2006. Report of The All-Party Parliamentary Inquiry into Antisemitism. London: The Stationery Office Limited.

American-Arab Anti-Discrimination Committee (ADC). 2004. “ADC Demands Merriam Webster Correct Definition of ‘anti-Semitism’”, March 16, 2004. http://www.adc.org/2009/11/adc-demands-merriam-webster-correct-definition-of-anti-semitism/.

Baker, P., Gabrielatos C., Khosravinik, M., Krzyzanowski, M., McEnery, T. & Wodak, R. 2008. “A useful methodological synergy? Combining critical discourse analysis and corpus linguistics to examine discourses of refugees and asylum seekers in the UK press”. Discourse & Society 19(3): 273-305.

Bank of English, 450 million word version, jointly owned by HarperCollins Publishers and the University of Birmingham.

Chambers's Twentieth Century Dictionary of the English Language: First edition, 1901; later editions 1964, 1988. Edinburgh and London.

Fairclough, N. 1995. Critical Discourse Analysis. London: Longman.

Fairclough, N. 2001. Language and Power. 2nd edition. London: Longman.

Finkelstein, N. G. 2005. Beyond Chutzpah: On the Misuse of Anti-Semitism and the Abuse of History. Berkeley and Los Angeles, Calif.: University of California Press.

Fisk, R. 2004. “A Warning to Those Who Dare to Criticize Israel in the Land of Free Speech. Another Case Study: Mary Robinson.” The Independent, 24th April 2004. http://www.informationclearinghouse.info/article6085.htm.

Forster, A. & Epstein, B. R. 1974. The New Anti-Semitism. New York, NY: McGraw-Hill.

Herzl, T. 1896. Der Judenstaat. Translated from the German by Sylvie D'Avigdor as The Jewish State; adapted from the edition published in 1946 by the American Zionist Emergency Council. http://www.mideastweb.org/jewishstate.pdf.

The Holy Bible, containing the Old and New Testaments. New Revised Standard Version. Mowbray, London: 1989.

Jewish Chronicle, “Fury as Leeds Uni bans lecture” and “An email that led to the ban”, 16th March 2007.

Mearsheimer, J. & Walt, S. 2006. “The Israel Lobby”, London Review of Books, 23rd March 2006.

Nunberg, Geoffrey. 2004. “Lexical Lessons; What the Good Book Says: Anti-Semitism, Loosely Defined”. New York Times, Sunday, April 11, 2004. http://www.nytimes.com/2004/04/11/weekinreview/lexical-lessons-what-the-good-book-says-anti-semitism-loosely-defined.html.

Oxford English Dictionary (OED). 1989. 2nd edition. Oxford: Oxford University Press.

Renouf, A., Kehoe, A. & Banerjee, J., 2007, “WebCorp: an integrated system for web text search”. Corpus Linguistics and the Web, ed. by C. Nesselhauf, M. Hundt & C. Biewer. Amsterdam: Rodopi.

Webster's New International Dictionary. 1890. First Edition. Springfield, Mass.

Webster's Third New International Dictionary, G. & C. Merriam Company, 1961. Publisher: Merriam-Webster, U.S. Preface by Editor in chief: Philip B. Gove. Springfield, Mass., June 1, 1961.