logo ELT Concourse teacher training
Concourse 2

Cognates and false friends

false friends

What are cognate words?

Cognates are words which have the same origin but exist in more than one language.
You may encounter the word paronym to refer to cognates but we'll stick with the usual term.

For example, the English word beer and the German word Bier both come from the same root.  One theory is that the words come from the Latin word biber [a drink] and another theory is that the words come from a very early German word for barley.  It really doesn't matter, even if it is quite interesting.  What is important is that the words have the same root and refer to the same thing.  This, of course, makes it easy for English speakers to learn the German word and vice versa.  The pronunciation, in both languages, is nearly identical although in English the final /r/ sound is usually droppeed.
However, there is also an English word bier which refers to the support for a coffin and is unconnected to the German word Bier.  (The translation of  the English bier in German is die Bahre and the words are cognates both coming from the same root and connected to the verb bear.)  The words bier in English and Bier in German are not cognates, although they look similar.

Words may be cognates in a range of languages but the meanings can drift apart over the centuries to the point at which the connection is hard to see unless you have a good etymological dictionary to hand.  For example, the English word clock (timepiece) is cognate with the German Glocke and the French cloche (both meaning bell) and the connection is clear but the words have diverged in meaning.


What cognates are not

Out here in cyberspace, there is a good deal of uninformed assertion, error and confusion about this matter.  Some careless, some ignorant and, just possibly, some deliberate.

Whatever you may read on the web, cognates are not:

  1. words which exists in the same language, with the same root but which mean different things.
    For example, in English, the words skirt and shirt both derive from the same word (the Old English scyrte) but they are not cognates.  The technical term is a doublet.  They derive from the same original source word (14th century, probably from Old Norse) by have drifted apart in meaning.
  2. words which are imported into one language from another.
    Partly for historical reasons, there are many hundreds of these in English taken from a variety of languages.
    For example, from the West African language Hausa, English has borrowed bogus, from Arabic, there is a huge range from admiral to zenith and zero, from Hindi, English has borrowed another large number, from avatar to veranda and from European languages there are many thousands more which began as borrowings but are now not recognised as such because the words are embedded in the language.
    Words may come from other languages via an intermediary language so a word borrowed from Arabic into Spanish (a common event) may then be borrowed again by English from Spanish.
    Such words can make it slightly easier for speakers of some languages to learn some words in English but the words are not cognates, they are, technically, loan words.
    There comes a time, of course, when what was a borrowing or loan word is no longer recognised as such.  For example, the estimate is that around 10,000 words entered English from Norman French from the mid-11th to the early 14th centuries and a further 7000 or so came from Parisian French during The Renaissance.  None of these words is now considered a borrowing because they are an accepted and unremarkable part of the English language.  Their existence does result in cognate words in both languages (and other Romance languages such as Spanish, Italian and Romanian).  Some of these words have retained the same meanings in English and another language and are helpful cognates, easy to understand and remember, but some have drifted apart and now constitute false friends.
  3. words which look the same but have different origins.
    For example, the German word Gift [poison] and the English word gift [present] may look the same but they are not closely connected and the English word hell (Hades) looks and sounds the same as the German word hell (light coloured) but they derive from wholly different roots.  Equally, the Japanese word for occur happens to be okoru but they, too, derive from very different sources.  The English word kitten, looks like the Tagalog word kuting and they both mean the same thing but are wholly unconnected.  Coincidences happen.
    Technically, these are called false cognates, not, please, false friends.

For a little more help in disentangling the terms, there is a language-question answer here which also considers a separate area known as pseudo-Anglicisms.


True and false friends


True friends help

It is sometimes difficult (and rarely necessary) to distinguish between a loan word and a cognate derived from the same source.  Indeed, it is difficult to identify the point at which a loan word ceases to be a borrowing and becomes part of the language as we saw above in the case of French.
For example, in Modern English we have many words derived from Latin directly but we also have words imported from Norman French which in turn were derived from Latin.  Most people would now call the words cognates, rather than loan words, whatever their route into English.
These are true friends to learners because the words not only look much the same they mean much the same.  They make reading comprehension, in particular, much easier.

The following lists could be extended very greatly, of course.  They are examples only.

Latinate words in English and some Romance languages:

English Spanish Italian French Portuguese Romanian
administer administrar amministrare administrer administrar administra
operation operación operazione opération operação operație
expression expresión espressione expression expressão expresie
communication comunicación comunicazione communication comunicação comunicare

Right across Europe a number of words derived from Greek will also be familiar.  Greek, too, is often the source of Latin words, incidentally.  In the following Italian is the only Romance language (i.e., derived ultimately from Latin) but the words will be similar in others (see last table).

English Greek Polish Italian German Danish
democracy δημοκρατία
demokracja democrazia Demokratie demokrati
athletics αθλητισμός
atletyka atletica Athletik atletik
theatre θέατρο
teatr teatro Theater teater
mathematics μαθηματικά
matematyka matematica Mathematik matematik

Greek still, of course, often forms the basis for coining new words, especially, in academic (ακαδημαϊκός [akademeyikos]) areas.

Among Germanic languages (such as English and most Scandinavian languages) there are also a number of commonalities but they are not usually as consistent as the ones above.  We get, for example, the English noun cook as Koch (German), kok (Dutch and Danish), kokk (Norwegian) etc.
We can also see cognate words across languages which share an ancestor.  For example, the names of days of the week in Germanic, Slavic and Romance languages are closely related within their groups (with a few oddities):

Germanic languages Slavic languages Romance languages
English German Swedish Polish Czech Slovak French Spanish Italian
Sunday Sonntag söndag niedziela neděle nedeľa dimanche domingo domenica
Monday Montag måndag poniedziałek pondělí pondelok lundi lunes lunedi
Tuesday Dienstag tisdag wtorek úterý utorok mardi martes martedì
Wednesday Mittwoch onsdag środa středa streda mercredi miércoles mercoledì
Thursday Donnerstag torsdag czwartek čtvrtek štvrtok jeudi jueves giovedi
Friday Freitag fredag piątek pátek piatok vendredi viernes venerdì
Saturday Samstag lördag sobota sobota sobota samedi sábado sabato

Numbers, of course, show similar patterns and are noticeably similar even between languages only distantly related with some numbers (such as six and seven) being clearly related across all nine languages:

Germanic languages Slavic languages Romance languages
English German Swedish Polish Czech Slovak French Spanish Italian
one eins en jeden jeden jedna un uno uno
two zwei två dwa dva dva deux dos due
three drei tre trzy tri tři trois tres tre
four vier fyra cztery štyri čtyři quatre cuatro quattro
five fünf fem pięć päť pět cinq cinco cinque
six sechs sex sześć šesť šest six seis sei
seven sieben sju siedem sedem sedm sept siete sette
eight acht åtta osiem osem osm huit ocho otto
nine neun nio dziewięć deväť devět neuf nueve nove
ten zehn tio dziesięć desať deset dix diez dieci

Family relationships are also fundamental concepts and are, unsurprisingly, similar across languages:

Germanic languages Slavic languages Romance languages
English German Swedish Polish Czech Slovak French Spanish Italian
mother Mutter mor matka matka matka mère madre madre
father Vater far ojciec otec otec père padre padre
sister Schwester syster siostra sestra sestra sœur hermana sorella
brother Bruder bror brat bratr brat frère hermano fratello

All these words derive ultimately from a language ancestral to almost all European languages known as Proto-Indo European and that language was also ancestral to Sanskrit, other South Asian languages and Persian among many others.
The presumed Proto-Indo European words were:
Father: pəter (which is related to the Sanskrit pitar, the Latin [and Greek] pater and the Old Persian pita)
Mother: mater (which is the root of many cognate words from South India to Ireland)
Sister: swesor (recognisable almost unchanged in Indo-European languages from the Sanskrit svasar to the Russian sestra)
Brother: bhrater (another almost unchanged word recognisable in the Sanskrit bhrátár, Old Persian brata, Greek phratér, Latin frater, Old Irish brathir, Welsh brawd etc.)

Problems with true friends

  1. Because the words look so similar, learners are tempted to pronounce them as they would in their first language(s).  This results in lots of incorrect stress and phoneme production.
  2. Because word formation, and especially the use of suffixes, works differently across languages, learners may mistake word class, confusing nouns with verbs and adjectives and so on.
  3. A word which is basically a true friend may have slightly different connotations or uses in particular registers.
    For example, Angst in German means worry or anxiety (both as nouns only) but does not carry the special psychological sense which the word has in English.  Similarly, the verb demander in French simply means to ask and carries none of the sense of insistence and authority that the verb demand carries in English.
  4. Many true friends in English for speakers of Latin-derived languages are stylistically inappropriate and too formal.  For example, while
        I administered an aspirin to him
    may communicate the sense, it would be more usual to use
        I gave him an aspirin.

False friends deceive

Cognates may have the same origin but often do not have the same meaning because languages develop in different ways.  For example, the word dish in English and the word Tisch [table] in German both derive from the Latin for a disc but they no longer mean anything like the same thing.  Fortunately, Tisch and dish look sufficiently different for learners of either language not to get too confused.  Some pairs of words are not so obliging.

In other cases, a word may be a loan word rather than a cognate but has developed a different sense in the language into which it was borrowed.  As we saw above, for example, the word Angst (German for worry or fear) has been borrowed in English but only in a special psychological sense.  The word smoking has been taken from English into French (and other languages) where it ends up meaning a kind of dinner jacket.

At other times, a word may derive from the same sources in a range of European languages but have either shifted in meaning in English or, as is often the case, remained true to the original meaning in English but drifted away from the original in other languages.
Here are some other examples of slowly shifting meanings of cognate words:

Whatever you may read on the web out here, these are false friends and are specifically and definitely not false cognates (see above for a definition of those).  Some false cognates may, in fact, be false friends but all false friends should not be called false cognates.

In the following, some words are cognates and others are loan words.  It doesn't matter for the purposes of helping learners recognise false friends and avoid embarrassing mistakes.

Take the following for example.  Can you fill in the gaps?  Click on the table when you have:

ff task

The list for all these languages can be considerably extended.  Moreover, related languages will often share characteristics so what is a false friend for an Italian speaker will probably (not certainly) be one for a French, Spanish, Portuguese or Romanian speaker, too.


Grammatical similarities and differences

Although the term cognate applies purely to lexis, there are similarities and deceptive differences between the grammar of related languages.
English is a West Germanic language (allegedly – there are those who prefer to call it a North Germanic language) and its grammar reflects its origin.  Although the language has been hugely enriched by borrowings from French, Latin, Greek and a range of other languages, its grammar remains recognisably Germanic.
There are obvious advantages for people whose first languages are Germanic when it comes to learning English but there are also some pitfalls which may be considered a form of grammatical or structural false friends.  Here are some common examples:

comparative and superlative forms
Cognates of the word longer in English and other European languages are instantly recognisable.  The word translates, for example, as:
    länger in German
    lengur in Icelandic
    längre in Swedish
    lenger in Norwegian
    langer in Dutch and Afrikaans
However, English cannot inflect the comparative form in the same way for longer words so, more interesting translates as:
    interessanter in German
    áhugaverðari in Icelandic
    interessanter in Dutch
and speakers of those languages may be tempted to produce *interestinger in English.
Speakers of Swedish and Norwegian will not because similar rules apply and the phrase is:
    mer intressant in Swedish
    mer interessant in Norwegian.
In the latter two cases, we have a form of interlanguage facilitation rather than interference.
modal auxiliary verbs
Across Germanic languages, these verbs are also recognisably derived from the same sources so must and can translate as:
    müssen and kann in German
    måste and kan in Swedish
    må and kan in Norwegian
    moet and kan in Dutch
However, the meanings are subtly different and the associated grammar can also be misleading.  For example:
    I must not
    I am forbidden to
in English, but means
    I don't have to
in many related languages.
tense forms
Superficially these look very similar so, for example, I have seen translates as:
    ich habe gesehen in German
    ég hef séð in Icelandic
    jag har sett in Swedish
    jeg har sett in Norwegian
    ik heb gezien in Dutch
However, the sense of the present perfect in English is rarely parallelled in other languages so most of these translations apply equally well to I saw.  In other words the relative or relational tense in English may be an absolute tense in other languages.
The same considerations apply to a range of other related languages in which a tense form may look similar but actually carries a distinctive meaning not parallelled in another language.
Conjunctions are usually ancient words which share roots in a range of related languages but have, over time, become distinctly different in meaning.  For example: the word because translates as:
    weil in German (cognate with the English while)
    därför att in Swedish (cognate with the English therefore)
    fordi in Norwegian (cognate with the English for that)
    omdat in Dutch (cognate with the English for that)
Similarly, the conjunction when in English translates as:
    wann in German (cognate with when)
    när in Swedish (cognate with near)
    når in Norwegian (cognate with near)
    wanneer in Dutch (cognate with when near)
and if translates confusingly as:
    wenn in German (cognate with when in English)
    om in Swedish (cognate with when)
    hvis in Norwegian (cognate with if)
    als in Dutch (cognate with as)
Apart from the difficulties of assigning the correct preposition to prepositional verbs such as depend on, stare at and so on, prepositions which are superficially derived from the same root are used very differently across a range of otherwise closely related languages so, for example, wait for in English translates as wait on in German (warten auf), Dutch (wachten op), Swedish (vänta på) and many other related languages.
The expression arrive at sometimes translates as arrive to, arrive at or arrive by, in related languages.
This is an obvious source of considerably error.
Other prepositions are more friendly so over as in over the wall translates as:
    über die Mauer in German
    yfir vegginn in Icelandic
    över väggen in Swedish
    over veggen in Norwegian
    over de muur in Dutch
Although articles are common to most Germanic languages, the way in which they are used is not always similar.  Many languages which are related to English in this regard can use an article as a pronoun which is not allowed in English and there are subtleties of use, such as when the definite article may be used with a proper noun, which vary between languages and catch out the unwary learner.

Related guides
the roots of English for more about the origins and development of English and its vocabulary which helps to explain why false friends occur at all
types of languages for a more technical guide to how languages vary and how they are related
interference and facilitation for a guide in the in-service area which considers how the nature of people's first language(s) can help or hinder the learning of English
false friends exercises some exercises for learners deliberately targeting false friends in a number of languages