Does language targeting stop my ads showing for foreign-language searches?

No. Language targeting decides who is eligible based on the languages Google believes a user understands — inferred from their interface language, account settings, browser, and behaviour across Google properties — not the language of the individual query. Google itself says the query language does not have to match the ad language: an English search can trigger a French ad if Google is confident the user understands French. So a multilingual user physically in your target market can type a query in another language and still be eligible for your ad, because they pass the language filter on who-they-are while the query text goes unfiltered. Query text is governed by keywords and negative keywords, not by the language setting, which is why foreign-language terms keep appearing in the report even with targeting set tightly.

Is Google really removing language targeting from Search campaigns?

Yes. Starting in September 2026 Google is removing the campaign-level language targeting setting from Search campaigns; ads will automatically match based on the language of the ad rather than a manual language selection. This matters for search-term hygiene because it takes away the one setting practitioners sometimes reached for to shape cross-language eligibility. After the change, negative keywords are effectively the only lever you control for blocking foreign-language queries that Google decides are relevant. If your account relied on a narrow language setting to hold back other-language traffic, plan to replace that control with negatives before the setting disappears.

How do I find foreign-language search terms in the report?

Pull the search terms report and scan for queries in scripts or words that are not your market's language — Cyrillic, Arabic, Chinese, Devanagari, or Latin-script words from another language. Sort by cost to surface the ones actually spending, and filter to zero conversions to confirm they are waste rather than legitimate multilingual customers. Non-Latin scripts are easy to spot; same-script foreign words (Spanish, French, Portuguese terms in an English account) are harder and usually reveal themselves as clusters of unfamiliar words with clicks and no conversions. Group what you find by language before you negate, because you will block each language as a set of its common words rather than as a single switch.

Can one negative keyword block an entire language?

No. Negative keywords match on the literal terms and their match-type variants, so there is no single negative that blocks a whole language. You block a language the way you block any query class: by enumerating its common words and adding them as negatives — articles, pronouns, and high-frequency nouns in that language, plus the specific foreign terms your report shows are triggering spend. For non-Latin scripts you add the actual foreign-script words as negatives, which Google Ads supports. This is the same enumeration problem as hand-built brand lists: you cannot list every word, so you target the high-frequency ones that catch most of the traffic and keep extending from the report.

Should foreign-language negatives go at account or campaign level?

Put languages you never serve on a shared or account-level negative list so every campaign inherits them — if you are an English-only business with no plans to serve Spanish or Russian, those languages' common words are true negatives everywhere. Keep language negatives at the campaign level when a language is served by one campaign but not another, which is common for accounts running separate campaigns per market or per language. The test is whether the language is unwanted across the whole account: if yes, lift it to a shared list; if one campaign legitimately targets that language, scope the negatives to the campaigns that should not serve it so you do not block the campaign that wants it.

Foreign-Language Search Terms: The Negatives to Add

Language targeting does not stop your ads from showing for searches in other languages, and that gap is about to matter more, because Google is removing manual language targeting from Search campaigns in September 2026 and letting ads auto-match on the language of the ad instead. Language targeting was never a query filter to begin with: it decides who is eligible based on the languages Google thinks a person understands, not the language of the words they typed. The query itself is query text, and query text is governed by your keywords and negative keywords. Once the manual setting is gone, negatives are effectively the only lever you control for foreign-language queries.

This post treats non-English (or non-market-language) queries as a distinct class of irrelevant search term, in the same way out-of-area place names are their own class. We will separate language targeting from language negation, cover the 2026 removal and what it changes, show how to find foreign-language terms in the report, catalogue the patterns worth negating first, and structure the negatives so a shared list does not block a market one campaign is deliberately serving. Google's own guidance is the anchor: the search terms report is where you find the actual queries that triggered your ads and turn them into negatives — foreign-language ones included.

Language targeting was never a query filter

The core misconception is that setting a campaign's language to English filters out non-English queries. It does not. Language targeting shapes eligibility by the languages Google believes a user understands, and Google infers that from a stack of signals: their interface language, account settings, browser language, and behaviour across Google properties. The language of the query is only one of those signals, and it is not decisive. As Google explains in its language targeting documentation, the language of the search does not have to match the language of the ad — a query in one language can trigger an ad in another when Google is confident the person understands both.

That is why a multilingual user standing squarely inside your target market can type a query in Spanish, Russian, or Arabic and still be eligible for your English ad. They pass the language filter on who they are, because Google has decided they understand English, and nothing in the campaign inspects the words of the query for language. The term serves, you pay, and it lands in your report looking like a targeting failure when it is really a keyword one — the same targeting-is-not-negation split behind location-name search terms that location targeting will not stop. The two problems rhyme: one is about where the searcher is, the other about which language they typed, and both are solved with negatives, not with a targeting setting.

The 2026 change: the one setting is going away

Whatever residual shaping the manual language setting gave you is disappearing. Google is ending manual language targeting in Search campaigns, with the campaign-level setting removed from September 2026 and ads matching automatically on the language of the ad copy itself. In practice this means you no longer declare "this campaign is for English speakers" as a lever; Google reads your ad's language and its own understanding of each user and decides eligibility. For most single-language accounts the day-to-day impact is small, but the control surface shrinks: the setting that let you narrow eligibility by language is simply not there anymore.

The consequence for search-term hygiene is direct. With the manual setting gone, the only thing standing between your budget and a foreign-language query that Google judges relevant is your negative keyword list. This is the same pattern as running broad match without a matching negative list — when the matching side loosens, the unpaired negative list is where the waste leaks in. Broad match and close variants already match on meaning rather than exact words, and Google's AI-driven query expansion can judge a foreign-language query semantically equivalent to your keyword. So the trend is toward more cross-language matching, not less, at exactly the moment the manual brake is being removed. Treat foreign-language negatives as a standing part of the list, not an occasional cleanup.

Finding foreign-language terms in the report

Start where every negative starts: the search terms report, sorted by cost and filtered to little or no conversions. Non-Latin scripts are the easy win — Cyrillic, Arabic, Chinese, Thai, Devanagari, and Greek queries stand out immediately against a Latin-script account, and a block of them spending with zero conversions is unambiguous waste. Pull those first, because they are the clearest and often the cheapest to eliminate completely. A handful of script-based negatives can close an entire vector of spend that a single-language business will never convert.

Same-script foreign languages are harder and where most practitioners miss spend. Spanish, French, Portuguese, Italian, and German queries share the Latin alphabet with English, so they do not jump out; they appear as clusters of unfamiliar words with clicks and no conversions. Read the report for words you do not recognise as your market's language rather than scanning only for odd characters, and group the hits by language as you go. This is the same discipline as any n-gram search-term analysis: you are looking for recurring tokens, and foreign function words — the equivalents of "the", "for", "how", and "near" — are the most reliable tokens to catch a whole language at once.

The language patterns worth negating first

Block a language the way you block any query class: by its common words, because no single negative covers an entire language. The highest-leverage additions are high-frequency function words — articles, prepositions, and pronouns — because almost every query in a language contains one. For Spanish that is words like "para", "como", "con", and "que"; for French "pour", "avec", "comment"; for German "und", "für", "wie". A short list of these per language catches the bulk of that language's traffic without your having to enumerate every noun. Add them as phrase or broad negatives so they match wherever the word appears in the query.

Then layer in the specific foreign terms your report actually shows spending, and for non-Latin scripts add the foreign-script words directly — Google Ads accepts negatives in other scripts. Be deliberate about which function words you pick, because a few overlap with your own language or with brand and product names: "con" is Spanish for "with" but also an English word, and "die" is a German article and an English verb. Where a word is ambiguous, prefer negating the specific foreign phrases from the report over the bare function word, so you do not accidentally throttle a legitimate English query. This is the same over-blocking risk that makes aggressive single-word negatives dangerous — the fix is to work from the report, not from a generic multilingual stop-word list.

Structuring language negatives across the account

Where a language negative lives decides how much work it saves you. Languages you will never serve belong on a shared or account-level list so every campaign inherits them: if you are an English-only business with no Spanish or Russian offering, those languages' function words are true negatives everywhere, and maintaining them in one place beats copying them into each campaign and missing one. This is the same logic as any deliberate negative keyword list structure across account, campaign, and shared levels: universal truths go high, specifics stay low.

Campaign-specific languages must stay at the campaign level, and this is where a careless shared list does real damage. Accounts that run a separate campaign per market or per language cannot put a language's words on a shared list, because doing so would block the very campaign built to serve that language. If one campaign targets Portuguese and another is English-only, the Portuguese function words are negatives for the English campaign but positives for the Portuguese one, so they belong scoped to the English campaign, not on a list both share. The test is simple: ask whether the language is unwanted across the entire account. If yes, lift it to a shared list; if any campaign legitimately serves it, keep the negatives local so you do not suppress the market you built a campaign to reach.

Cross-check before you negate a wave of languages

Confirm you actually serve a single language before mass-negating others, because the cost of a mistake here is invisible: you will not see the impressions you suppressed. Check whether any current or planned campaign targets a bilingual or immigrant audience in your market — a business in a city with a large second-language community may genuinely want those queries, and blanket-negating the language would quietly cut real demand. Where you serve multiple languages, negate only the languages you do not serve, and keep each served language's words off the shared list. When in doubt, negate the specific foreign phrases the report shows converting nothing, rather than the whole language, and widen only once you are sure.

Then fold the check into your routine. Once the manual language setting is gone, foreign-language terms behave like any other class of waste: they surface in the report, you group and negate them, and you revisit when a genuinely new language appears. Run it as part of a standing search hygiene audit, and treat a rising count of other-language queries as a prompt to extend the negatives rather than to hunt for a targeting setting that no longer exists. With the served languages protected and the unwanted ones enumerated from the report, cross-language waste becomes a solved, low-maintenance line item instead of a recurring leak.

For the sibling class of irrelevant terms and the same targeting-is-not-negation logic, see location-name search terms and why targeting will not stop them. For the match-type mechanics that make cross-language leakage worse, see broad match without a paired negative list, and for where the negatives themselves belong, see negative keyword list structure across account, campaign, and shared.