Uli Kozok

Waltzing Matilda vs. All Blacks:

Quantifying Interest in Asian Languages

Many erroneously call the Internet the first democratic medium. Yet, it hardly qualifies as a medium and by no stretch of terminology is it democratic. The average user of the Internet is thirty years young, with academic background and high income. The percentage of the educated and the well-to-do among the users of the Web is three times as high as their proportion in the population. As far as content is concerned, the Internet cannot be currently defined as a medium; at best in can be described as a rather disordered library. Despite its limitations, it can be cautiously used to quantify interests or preferences in different countries or cultures although the results will not be representative for the whole population but rather for the typical Internet user as described above.

According to our expectations we will find many more Internet sites on Maori culture in New Zealand than in Australia. It is hardly surprising that a search term such as "dairy board" or "Waitangi" will create many more hits at Internet sites hosted by a New Zealand Internet provider with the domain suffix "nz" than at Australian-based Internet sites. A word- or phrase-search for Australian cultural icons on the other hand should yield higher on Australian based sites with the suffix "au". Search-words that are culturally unspecific, and especially those concerning frequently used terms, should then, as one might expect, yield 'hits' proportionally to the population of the two compared countries - in our case New Zealand and Australia with a population of 0.2:1. [1].

An Internet search performed by a popular search engine such as Google <www.google.com> for common words or phrases such as "language", "Toyota Corolla, or "Chardonnay" should thus yield about five times as many hits at "au" than at "nz" sites. If a search word or phrase has a "nz–"au" ratio of 1:5 hits, it has been assigned a Popularity Index (PI) of 100. A search conducted for 25 very common words (resulting in a minimum of 5,000 hits at "nz" sites) resulted at 2,139,470 hits at New Zealand, and 10,050,700 hits at Australian sites with a PI of 94. The typical range for very common search terms ranges from a popularity index of about 50 to 150 (Television [57], radio, [75], car [100] Internet [117], food [147]).

The higher the PI the more popular is that search term in New Zealand sites relative to Australian based sites.

	"nz" sites	"au" sites	PI
Dairy board	2530	208	6100
Waitangi	19900	1720	5800
All Blacks	3960	2730	725
Typical range of cultural unspecific, and very common words or phrases			150
			50
Margaret Preston	315	3890	40
Waltzing Matilda	66	3330	10
Alice Springs	369	69000	3
Table 1 Google Hits and Popularity Index of NZ and OZ cultural specific search terms

Table 1 shows clearly that culturally unspecific terms have a PI close to 100 whereas terms closely associated with New Zealand have a high popularity index. Australian icons score high in Australia, and relatively very low in New Zealand. The All Blacks have thus a very high, and Waltzing Matilda has a very low PI since the point of view taken in this study is a New Zealand point of view:

Each search engines operates by using parameters that are constantly changed. It is thus not surprising that we have considerable differences when using different search engines such as Hotbot and Google. The difference are, however, not big enough as to distort the overall picture. I have used Google to quantify the results since Google is a meta search engine (operating by using a number of independent search engines and displaying the total computed hits), and also because Google gives exact numbers whereas Hotbot tends to display vague numbers such as "more than 500" or "less than 200". Google is also widely acknowledged as one of the most powerful and accurate search engines. As more and more Internet pages are added to the Internet the PI also fluctuates constantly. This study was for controlling purposes repeated several weeks after the original data was obtained, and it was noted that there were some, but in most cases insignificant deviations in the PI.

When it is possible to detect certain trends in popularity by simply counting the websites where these search words or phrases appear, then it should also be possible to quantify interest in foreign languages.

A search for "language" results in 389,000 hits at Australian sites and 84,700 hits at New Zealand sites, which is quite close to the expected 5:1 distribution (PI 110).

	"nz"	"au"	PI
samoan language	130	27	2407
tongan language	33	35	471
fijian language	30	64	234
Persian language	31	88	176
latin language	80	269	149
hawaiian language	11	39	141
dutch language	56	210	133
german language	563	2480	114
russian language	109	650	84
spanish language	225	1470	77
japanese language	692	4570	76
polish language	21	157	67
danish language	9	68	66
urdu language	5	43	58
hebrew language	55	494	56
french language	310	2840	55
chinese language	384	3610	53
korean language	81	883	46
greek language	94	1040	45
Malay language	14	187	37
tamil language	7	95	37
hindi language	10	146	34
portuguese language	17	264	32
Italian language	92	1560	29
thai language	27	466	29
turkish language	9	205	22
Arabic language	23	655	18
indonesian language	55	2620	10
vietnamese language	8	573	7
Table 2 Google Hits and PI of foreign languages

Among the high interest languages of New Zealand is, of course, Maori. Reference to "Maori language" yields 612 hits in Australian, and 3730 hits at New Zealand sites (PI 3050). Maori is followed by Samoan, Tongan, Fijian, and Hawaiian, indicating that New Zealand has firmly associated herself with the other Polynesian nations.

Compared to Australia, the interest in Asian languages in New Zealand, seems to be relative low. Of all Asian languages, Japanese has the highest PI (77). All other Asian languages (highlighted in Table 2) score much lower with Indonesian and Vietnamese at the very end of the list.

In examining the absolute numbers the picture looks not that much better for Indonesian either. Of the 29 languages examined, including 14 Indo-European, nine "Asian" languages (excluding the Indo-European languages of the subcontinent, i.e. Hindi, Persian, and Urdu), four Polynesian, and two Semitic languages, Indonesian yields 55 hits which made it the 14^th most popular language in New Zealand - a striking contrast to Australia where Indonesian is the fourth most Internet-popular language. In New Zealand Indonesian and Malay together would just barely outnumber Dutch as the 11^th most popular language. But if you are not one of the top-tens, it is hard to survive in a small country. The teaching of both Dutch and Indonesian at tertiary level in New Zealand was terminated after the University of Auckland closed both its Dutch and Indonesian language program in 2001. Given the accelerated rate in which New Zealand universities get rid of their language programs Latin, Korean, Greek, and Russian are likely to follow in due time.