foreign google searches contain also other language results

for example a dutch only query giving english, german and uk results

[janjaap]

If I search on google with dutch only resutls I am also getting foreign language results.

for example I search for the word google:

[google.nl...]

the url shows that I only want to have websites with dutch language but I get the following results:

1. correct (www.google.nl)
2. correct (www.google.be)
3. correct (.....agina.nl)
4. incorrect (.....hosting.de)
5. incorrect (.....osinga.com)
6. incorrect (news.com.com)
7. incorrect (.....ciousreport.com)
8. incorrect (.....uimp.com)
9. incorrect (.....ogenix.org)
10. incorrect (.....inesswire.com)

Does anyone else see that too?

[Brett_Tabke]

interesting search.

[tenerifejim]

When searching for place names which are the same across numerous languages (e.g. Paris) even though I am searching from .co.uk with english specified:

[google.co.uk...]

Strangely, if I search with no language preferrence and only request pages from the UK I get better results, but still a foreign language site at no. 6.

[google.co.uk...]

This is a really pain for me because my main site is a place name that is similar accross many places but I only want to appeal to english speaking people (my site is in English). I hope I don't appear in SERP for Germans or French.

[bird]

That's exactly the reason why people (including Google...) should be using the lang="xx" attribute in their html tags. Automatic language detection is tricky.

[tenerifejim]

I agree with your suggestion. But could you explain to me exactly what is so difficult about recognising French, German, English, Italian and Spanish as distinct languages?

I google can't determine the language of a page, then what chance has it got of doing all the other matching it does (word stemming, etc.)

I don't think this is an issue with google's ability to determine the language a page is written in. I think it is something else.

[vitaplease]

In general I thought Google became better over the last year or so in detecting languages with regular words existing as such, in different languages. In the past many Danish and German results could turn up on Dutch searches.

Google seems to mainly be troubled with searches of non-dictionary type of words?

Lets say searches for Sony, Google, Intel...

Yahoo seems to do a better job on serving Dutch only pages for the query on "Google".
Alltheweb on the otherhand is a mess.

[Jeroen van de Wiel]

I noticed the same thing in the past week and it seems that the language search (probably especially in Dutch) is extremely buggy at the moment. Over 50% of the so called Dutch results are in foreign languages which makes this option for the Netherlands almost impossible to use. I waited with posting since sometimes bad things are gone themselves, but something more seems to be going on?

Maybe GoogleGuy can help with what is going on?

[ambiorix_rb]

== Yahoo seems to do a better job on serving Dutch only pages for the query on "Google". ==

I never thought I would write this but I this is true. Searching for dutch kw in Google.be (belgian pages) the results are really poor (very poor).
Sometimes we have .com, .jp(?) and even .ru(? cyrillic pages)

When searching on Yahoo! I have only true belgian sites (no strangs domains) and also no PDF results.
In google.be, sometimes 75% of the SERP are PDF files from way back in time...

Just my 2 cts
Ambiorix

[tenerifejim]

I am planning to launch areas of my english site into German, French, Spanish and Swedish soon, with dutch, polish and czech to follow. This could certainly affect how much I concentrate on specific languages. IS there any feed back out there for French and Spanish results? How good are they?

[Leosghost]

from what I've seen the French results are pretty much as normal ..however that said ..so many words in French have been ousted by their English counterparts ( inspite of the efforts of "l'acadamie " et Mr Toubon )..that they always were a touch "Franglais"...
The only way you can really tell in any country market is to restrict search to only say French webpages or whatever and then test using a word which does exist in more than just that language ..which does not however have the same meaning at all in the other languages ....in French one such word ( there are thousands and thousands ) is "permanance" ...in English it usually refers to the "lightfastness of pigments" ...in French it means " occasional manning or wommaning : ) of an office or similar ) ...if you see what I mean ...IE ..when you see that some office has a "permanance" ..it actually means that ..no they don't ..they are only there sometimes each week or month ....( Ray Bradbury once said that France was the "Martian Embassy on Earth"! )....

This word when tested gives me back mainly French uses of the word as included in serps so I presume that google is currently working as "She" should in France ....

[encyclo]

The same thing happens when searching for French-language only pages on Google.ca - what's more, when an English-language page is offered in the SERPS, Google helpfully offers to translate it from English into French - which proves that (at least in some cases) they can correctly identify the document's language, but that their language filter is broken.

<added>Oh, and welcome to WebmasterWorld, janjaap! (where are our manners ;)</added>

[Leosghost]

Yes ..sorry ..welcome to WebmasterWorld janjaap...( Hey and his first reply was from the Boss ) ..this guy will go far methinks ..

Encyclo ...? does the .ca center return "metro-french" or "quebec" definitions for French words ...?

I know the "Larrousse" editions are not thes ame for the two versions of French as we have both in the house ..I prefer the quebecois version even tho I think that celine and singing is an oxymoron ...

[Herenvardo]

IS there any feed back out there for French and Spanish results?

I'm in Spain and frequently make searches for Spanish pages. about 30% of results are non-Spanish (normally italian and portuguese)

Herenvardö

[troels nybo nielsen]

Some hours ago I checked Danish results for a non-Danish name that is a primary keyword for two of my Danish websites.

#4 in the SERP was in English
#5 was in German
#8 and #9 were in Norwegian
#11 was in French
#17 was a links page on a Croatian website and lots of link titles in many different languages.

There is nothing new in having Norwegian results in Danish SERPS. Search engines often have significant difficulties in distinguishing between Norwegian and Danish texts and I don't think that it would be fair to criticize them too much for that.

Nor is it seldom to see English results. But I believe that this is the first time that I have seen a German result anywhere near the top in a Danish search.

I'm not so sure that this was a result of recent changes in Google's algo, though. In fact it may partly be my own "fault". This German result was shown only with title and URL, indicating that the search term was only found in links. And the cache showed an empty frame. This ought to make it a bit hard for Google to accurately decide the language of the page. And one of the incoming links to it actually is from a Danish page of mine.

I believe that if you carefully investigate those pages that appear in "wrong" languages you will often be able to find a sensible explanation why they do.

Now I just did the same search again. And top 20 is completely Danish apart from a Norwegian page at #8 and the French page now being #10.

I have seen on some other searches that Google seem to have shuffled things about. Perhaps they have done some cleaning?

[ThomasB]

I guess the language of the linking site is the solution. Maybe G should reduce the value of links ;)