Wiktionary:Etymology

General
Etymology is the study of the origins of words. The vocabularies of modern languages come from a variety of different sources: some have evolved from older words, others have been borrowed from foreign languages, and some have been named from people, developed from initialisms, or even have been deliberately invented by a certain author. Etymology sections in entries of the English-language Wiktionary provide factual information about the way a word has entered the language and usually some sense of its semantic development.

Brief
Etymology sections should not be too verbose, particularly because they appear before the definitions; usually a simple list of previous forms is all that is required. They should be based on reliable sources but should not be copied verbatim from another source.

Some words may also benefit from further details, such as cognate words in related languages, or some illustrative comments.

There is currently no standard for longer discussions of etymology.

Lemma
Include the etymology on the main entry (the lemma), even if historically it derived from another form, such as by back-formation.

Folk etymologies
Folk etymologies should not usually be discussed in entries. This can however be discussed where the folk etymology is widespread. However priority should be given to the 'correct' etymology, with the folk etymology being kept brief to avoid unnecessarily lengthening the etymology section. Unreasonably long material may be moved to the talk page, where it can be discussed at length.

Surface etymologies
Etymologies trace the historical development of words, not simply an analysis of their current (“surface”) forms. For example, comes from Ancient Greek  (via Latin), though its surface form can be analyzed as  +, as the components are valid English combining forms. Conversely, does not come from an Ancient Greek term, but is rather a, coined c. 1800.

Analyses of surface forms are of value, but do not replace and should not be confused with an account of historical development. For example, the entry for should not read *“From  + ”, but should instead read “From Latin, …. Surface analysis + .”

Phrases, compounds, acronyms, and abbreviations
For a term that is composed of base words separated by spaces or hyphens, do not add an etymology that just notes the base words. This can be better shown by wikilinking the term in the inflection line. The English term computer language, for instance currently has no etymology, and the base words are automatically linked to with the headword line template. Similarly, the etymology of acronyms or abbreviations is simply the definition, and no separate etymology is necessary.

Conversely, for compounds – a single word without spaces or hyphens, such as – a brief etymology section using  is useful, as wikilinking the components in the headline of the entry does not distinguish the components (it would just appear to be a single clickable link).

However, if some etymology would prove useful, it should be included – for example, history of usage or coinage, such as (coined during WWII for the chaos of war), or explanations of set phrases or idioms such as  (hangover cure, from folk remedy for rabies), or the origin of a proverb or word that can be traced to a particular source or sources such as.

See also Compounds.

Etymology jargon
Some words have conventional usage in etymology:
 * from
 * Using a bare “from” denotes a single step, with no intermediate steps – a direct descendant or borrowing – as in: “From French, from Latin, from Proto-Italic, from Proto-Indo-European.”


 * ultimately from
 * Using “ultimately from” indicates that some intermediate steps have been elided when there are many intermediary steps or the specific languages are unclear. For example, “Ultimately from Proto-Indo-European” (but by some other languages in between), or at 🇨🇬, “Ultimately from 🇨🇬” (through an unclear intermediary).


 * akin / related
 * The term “akin” is used to indicate an attested word that is presumed to be etymologically related, when the ultimate etymon is not attested. This is used particularly for proto-languages, for language groups, and for unattested terms in attested languages.


 * For example, in tracing an English word back to Proto-Indo-European (which is not attested), presumed cognates of its Old English parent word can be referred to with “from Old English X, akin to Old High German Y, Latin Z, etc.”
 * Similarly, if a word can be traced back to an indeterminate Germanic language, one can give examples of related attested words, but not state a specific etymon (because unknown), writing for instance “of Germanic origin; akin to Old Saxon X” (but might be from Old Frisian or another language).
 * “Akin” can also be used when the specific etymon is not attested in an otherwise attested language, for example: “: ultimately from Latin + base akin to ” (but * is not attested).
 * "Akin" is a weaker claim than "cognate to". The former only implies relationship in some, possibly so far undetermined fashion, while the latter is commonly understood to imply descent from a common ancestor.

Layout: Borrowed or inherited words
There are numerous types of word origins, including borrowing and word formation mechanisms, followed by processes of lexical change, notably sound change and semantic change. These can be formatted in conventional ways, as detailed below.

From vs symbol
There are two ways to denote the sequence of derivations in an etymology section. Historically, some editors used the word “from” to separate ancestors, while others used the algebraic “<”. (The symbol “<” implies an arrow that points in the direction of language change.) A 2011 vote did not reach a consensus on a preferred form, but since then, the number of entries using “<” has declined to the point that “from” is Wiktionary's de facto standard. However, “<” is still found, especially to show the progression of intermediate forms:
 * (at ), from an earlier *leite < *laite, from , from , from . Compare 🇨🇬.

Inherited words
A significant category of words in a language are the so-called ‘native’ or ‘inherited’ words; in some languages, but not all, they form the majority of words. This means that they have developed from an earlier form of the language which may or may not have gone by the same name. Some of these ancestor-languages were written down and are well-attested, but others are not. For example, French, Spanish, Italian, Romanian and Portuguese all developed from Latin. The French word, for instance, and the Spanish word both evolved from the Latin word  (they are cognates). They were not borrowed from Latin; the Latin language evolved naturally in different areas into the different forms.

The ancestors of English are, in order: Middle English, Old English, Proto-Germanic, and Proto-Indo-European, and native words are those that came from these ancestors, without at any stage being borrowed from a different language, nor by being borrowed from an ancestor at a later time.

One should show the complete sequence of ancestors if possible, not just the immediately preceding form.

To specify the source language from which a term was inherited and to link to the original term, use. See Templates, below, for details. If an ancestor had the same meaning as the descendant, it does not require a gloss in, but if the meaning changed, one should gloss the ancestor, showing the earlier meaning.

For native words, you can show the sequence of ancestors as follow, as in :

===Etymology=== From, from , from, from.
 * From, from , from , from.

The initial ancestor can be prefaced by “From”, assuming it is different from the current form, but some editors prefer not to preface it by “From”.

Even if the current form is identical to an earlier form, the same format should be used, so that identically spelled words in earlier language can be linked. For example, is identical in spelling to its immediate Middle English ancestor, giving the following format: ===Etymology=== From, from.
 * From, from.

Reconstructed terms
In some cases, the ancestor terms were not recorded. The most recent ancestor language to English, German, Swedish and Dutch, which was spoken around the same time as Classical Latin, was not written down. We call it Proto-Germanic, because it developed into the various Germanic languages, of which English is one. Many words from this language can be inferred with great confidence by comparing the surviving forms in daughter languages.

Some terms may be unattested in an otherwise known and attested language. For example, many Latin terms were newly developed and came to be used by the common people of the late Roman Empire, yet scribes did not write them down as they were colloquial rather than literary. We call these terms Vulgar Latin and they form the basis of the modern Romance languages.

Because reconstructed terms are not attested, there are different rules for them. They must always be preceded by an asterisk * to indicate that they are reconstructed, and entries for reconstructed terms are not allowed in the main namespace. Adding * before a reconstructed term ensures that the links point to the right place.

Some examples:

yields:

yields:

The inclusion of cognate words in related languages is particularly useful for inherited words, since they show how the same original form has developed in different daughter languages.

In the entry :

===Etymology=== , from. Cognate with 🇨🇬, 🇨🇬, 🇨🇬.
 * , from . Cognate with 🇨🇬, 🇨🇬, 🇨🇬.

Borrowings
Some words have been borrowed from other languages, either because of a historical occupation or co-existence, or simply through exposure to other languages. For example, the English word is borrowed from Latin, which itself was borrowed from Ancient Greek. Borrowings can be ancient or recent. When words are first borrowed into a language they may still ‘seem’ foreign; examples in English include or. After a while they become more naturalised—like French borrowings from the last century such as or. Eventually they seem completely native, such as or  (borrowed from Old Norse and Latin respectively).

Beware to differentiate Ancient Greek, using the language code grc for Ancient Greek, as in , not the code el, which is for Modern Greek.

Key waves of borrowings into English are from (non), Anglo-Norman (xno), and  (frm) in the 14th century. See English language: Word origins for discussion.

Many words in English are also ultimately of French (fr), (la; see ), and  (<tt>grc</tt>) origin, but often not directly.

While there are some borrowing from modern French, such as, many terms are instead from Middle French, or from (<tt>fro</tt>) via Anglo-Norman. Many terms of Latin origin entered English via French, while many terms of Ancient Greek origin entered English via Latin or French or both; see also Stages of Latin, below. Further, many terms with Latin or Ancient Greek etymologies are not borrowings from ancient languages, but instead are s – modern coinages based on nativized s. For example, is not borrowed from Ancient Greek, but is coined from  +.

In addition, in modern times especially, English has borrowed from a great many languages.

Stages of Latin
Further, it is useful to differentiate which stage of Latin a borrowing is from –, followed by  (written;   3rd c.–6th c.) and  (spoken;   3rd c.–9th c.),  (  6th c.–c. 1500), and  (  c.1500–now). These non-ISO codes are used in and other etymology templates; for example,  contains: which yields:

Note also that Vulgar Latin terms are generally not attested, hence the reconstructed terms should be linked to with an asterisk, as in : which yields:

See Dialects for details on these codes.

Differentiate borrowings
If any step of a word’s history is a borrowing, this step should be flagged as such; in English, any word not inherited from Middle English, from Old English, from Proto-Germanic, from Proto-Indo-European, is at some stage a borrowing.

Languages may borrow from an ancestor at a later date: for example, the two Spanish words and  both come from Latin, but the former was a natural development (hence ‘native’), whereas the latter was borrowed back into Spanish much later (in the fifteenth century in this case).

Example
To flag a borrowing, use or the shortcut ; for instance, in the entry :

===Etymology=== Borrowed from, from. Compare.
 * Borrowed from, from . Compare.

Borrowed forms
A form of a word may be borrowed, in which case one should say “From Xus, form of X”, where Xus is the form, and X is the lemma.

See Lemmas for which form of a verb is the lemma in various languages; notably for Latin, the lemma is the first, the first person present indicative, such as , rather than the infinitive (such as ), which is the lemma in English.

Beware that a form may be borrowed, and then other forms created by regular formation or back-formation, while in other cases different forms may be borrowed independently, as in the below example: was borrowed into English from Latin, derived from Latin , while this latter was also borrowed into English as  – stimulus/stimulate are not formed from each other in English by a regular rule or back-formation.

Example
In the entry :

From, perfect passive participle of , from.
 * From, perfect passive participle of , from.

Modern form
When listing a borrowing from an old language, it is useful to list the modern inherited cognate in daughter languages of the original language. For example, in English, one may write:
 * From (Icelandic, Swedish , Danish , Norwegian Bokmål , Norwegian Nynorsk ).

Displaced native words
Borrowings may displace a native word, rather than providing a new word; this may also occur with new coinages. If this is so, one may list the displaced word in the etymology section, provided that this does not make the etymology unnecessarily long. Examples include English (replaced ) and French  (replaced ).

Inflected forms
For words that are not considered separate lemmas, but rather inflected forms of another word, etymologies are not usually added. This includes plural forms, inflected verb forms, case forms and so on. In such words, the etymology is usually implicit in the definition of the form, and should not be stated separately. However, if the formation is irregular in some way (such as from ), an etymology might be useful. Additionally, the template nonlemma is often used in entries with multiple etymology sections in which one or more sections is a lemma, to indicate that a more complete etymology is not missing from non-lemma sections.

Affixation and compounds
If a word is formed by a regular rule, such as adding an affix, it is preferred not to repeat the complete details of the base word’s origin on the page for the derived or inflected form: simply show the rule, and leave the more informative etymology for the lemma of the base word.

The templates and  (with the arguments <tt>t1</tt> and <tt>t2</tt> for translation if components are not clear) are useful for this, and place entries into the correct categories (“words prefixed/suffixed with …”), as in the following entry for : ===Etymology=== From.
 * From.

A compound word is a word composed of two or more smaller words, but used as a single unit, like science fiction or website. For these, the etymology can simply be.

Back-formations
Conversely, words that look like a regular formation can have the formation reversed (especially, removing apparent affixes), yielding a new word. This is called back-formation, and the template helps here.

Not to be confused with clipping, which is just a shortening of a word, not the undoing of a formation, and does not change the meaning or part of speech.

Note that back-formations are generally the lemma entry, and should have the more informative etymology, rather than relegating the earlier etymology to the etymon.

Examples
In the entry :

===Etymology===
 * Back-formation from.

Blends (portmanteau words)
A blend or portmanteau word is a word which was originally formed by combining two other words. For example, is a blend of  and.

Examples
In the entry :

===Etymology=== .

Coined expressions
In some historically recent cases where words have been deliberately created, we may be able to give details of where and by whom this was done. Where possible, the reasoning behind the coinage should be suggested, however note that this will properly be conjectural unless it has been documented by the word’s original creator.

If the original coinage is attested, common practice is to include the relevant quotation in the etymology, and link to a source, if possible, as in serendipity or portmanteau word.

Examples
In the entry :
 * . Ostensibly from.

In the entry :
 * in Jabberwocky, apparently as a.

Calques
For calques or loan translations it is necessary to provide the source language out of which the lexeme, compound or a phrase has been calqued. Sometimes the exact source of calque cannot be established due to its spread among several languages, in which case several notable examples should be listed. The template should be used, which automatically categorizes the term as

Examples
In the entry :

Etymology
, a.
 * , a calque of.

Abbreviations
For acronyms, initialisms, and abbreviations that have a foreign origin, such as qv or cf, an etymology section should describe the foreign expansion. For native acronyms, initialisms, and abbreviations (including single letters used symbolically), the definitions are simply the expanded terms. A separate etymological section for each meaning is therefore unnecessary, and multiple senses should be grouped together under the same L3 header. A generic etymology section can be used to separate abbreviation definitions from non-abbreviated definitions. See A for an example.

Layout: Phrases
For phrases, which are terms that consist of several words, no special etymology is normally needed. For convenience, the individual terms in the headword are often linked to their respective entries, like so:

Note that if a phrase is actually a compound word, an etymology with (see above) should also be provided to show that the term is a compound, and to add the term to the appropriate category.

For phrases that have more complicated origins, an etymology may be useful. This applies in particular to idiomatic phrases that cannot be interpreted literally by the sum of their parts, such as. For idiomatic phrases in languages other than English, the etymology can be used to provide the literal translation of the phrase.

Further details
In addition to etymology, one may provide years and location of origin, cognates, and glosses in the etymology section.

Year and location
In addition to which etymon a word comes from, it is useful to state when and where the word came into use – specifically when it is first attested, and location of origin, if known (e.g., US vs. UK, region). For example, is of US 1920s origin. This can and should be brief – “US 1920s” is preferred to “First attested in the United States in the 1920s period”. Further, one should state (briefly) when (and how) other senses came into use, if they differ from earlier meanings.

Cognates
Cognates are not strictly part of the etymology of a word, but can provide useful context, as well as serve as a mnemonic device.

The inclusion of cognate words is allowed only for inherited words, deriving from the same etymon in the ancestor languages, since they show how the same original form has developed in different daughter languages. This is especially useful for words whose ancestor form is not attested, and where regular sound correspondences can be observed.

For borrowed words, one may instead list the modern form (inherited cognate) in daughter languages of the original language. For example, in English, one may write: “From (Icelandic , Swedish , Danish , Norwegian Bokmål , Norwegian Nynorsk )”.

Cognates can be listed at the end of the etymology as: “Cognate to lang. term.” One may use the template.

Care should be taken however: These are general guidelines and individual language/language-family policies take precedence over them.
 * 1) Not to overburden the etymology section with too many cognates. One example from a major branch of the immediate ancestor should suffice, with at most 4-5 cognates listed. Users can always look up more cognates descending from the attested or reconstructed form in the corresponding etymon's ====Descendants==== section, or the page of a reconstructed ancestor.
 * 2) Not to significantly mix cognates diachronically, by listing cognates of modern languages in the etymology section of ancient languages, or by listing cognates of ancient languages in the etymology section of modern languages. Ancient languages should thus prefer ancient cognates, and modern languages should prefer modern cognates. E.g. in case of the Indo-European language family that means Ancient Greek, Latin, Old Church Slavonic, Sanskrit/Avestan/Old Persian, Lithuanian/Old Prussian, Gothic, Old Irish, Tocharian, Old Anatolian (Hittite, Luwian, Palaic) and Old Armenian. Exceptions are "single-language" families (e.g. Armenian, Albanian), and cases where there is no ancient but only modern cognate attested, usually occurring in case of languages that were attested relatively lately (e.g. Lithuanian, Albanian) and these are allowed to list ancient languages as cognates.

Additional notes for the Indo-European languages:
 * Ancient languages of large sub-branches (Indo-Iranian, Germanic, Balto-Slavic) should list both their ancient language cognates of the same branch, and the ancient Indo-European cognates.
 * In case of ancient languages, Old English (Anglo-Saxon) is an exception and should always be listed if it is a cognate, including the modern-English reflex in parentheses.
 * Special care should be taken to always mutually link cognates of "classical" ancient Indo-European languages: Sanskrit (Vedic), Latin and Ancient Greek.

Glosses
In some cases where the semantic development is not obvious, some explanatory comments may be useful. The more concise and efficient, the better.

Examples
In the entry :
 * From the stage adaptation of George du Maurier’s novel Trilby, in which such hats were worn.

In the entry :

Borrowed from, from , from , an ancient Greek city in southeastern Italy noted for the luxurious, pleasure-seeking habits of many of its inhabitants.
 * Borrowed from, from , from , an ancient Greek city in southeastern Italy noted for the luxurious, pleasure-seeking habits of many of its inhabitants.

Descendants
Complementary to etymology (going backwards) is descent and derivation (going forwards): as per WT:ELE, please link back descended terms in the “Descendants” L4 heading of the ancestor term, and likewise for derived terms is the “Derived terms” L4 heading: descendants are terms either inherited or borrowed into another language, while derived terms are terms in the same language which derive from a given term.

Closely etymologically related terms in the same language should be listed instead at “Related terms”, and there should be links both ways, as this is a sibling relationship; related terms in other languages are instead handled as cognates in the etymology section, or as descendants of a common ancestor term.

Proposals for the format of the descendants section are discussed at Wiktionary talk:Descendants, and a specific format policy is at WT:Latin: Descendants, but as of this writing, there is no detailed general policy. You may use to create a link to the correct language. Narrowly, you may wish to distinguish inherited terms from borrowings by suffixing the latter with “(borrowed)”, and list descendants at the form from which they are descended (rather than the lemma), but this is at discretion.

Etymology language templates
There are several templates commonly used for etymologies. All templates use the two- or three-letter language codes listed in List of languages.

The templates (short for ),  (short for ), and  (short for ) should be used whenever possible. The target language comes first, then the source language and the term. For instance,.

is used when a word is inherited from an ancestor language with regular sound changes and no change in form; is used for words borrowed from a language that is not an ancestor; and  is used in all other cases. These templates categorize a term in the proper category: for instance, Category:English terms inherited from Old English.

A comprehensive example for a native English word is ; the language codes <tt>enm</tt> and <tt>ang</tt> represent Middle English and Old English: ===Etymology=== From, from , from, from.
 * From, from , from , from.

For cognates, one can use, which takes the same parameters as , but displays the language name.

Other templates
Other useful templates are, and , for flagging stubs or disputes. As many entries lack etymology, this is most useful if there is a partial etymology; including it for all entries lacking etymology would be distracting.

Where a term originates in a foreign, but undetermined language, use. In cases where an etymology is reliably identified as unknown, may be used (note this can be used for native-born terms unlike the previous template).

Structure of etymology categories
Following the Wiktionary convention for parts of speech categories, each language has its own root etymology category. For example, the root etymology category for Scots language is Category:Scots terms by etymology.

Similarly, for each of the derivations categories (e.g. Category:Terms derived from Old English) the corresponding category for example for Scots language would be Category:Scots terms derived from Old English. The template should be included for all of these categories, and an appropriate subpage created (if necessary). (The previous template,, is deprecated.)

The etymology categories are inclusive: they include all terms that trace their roots to the source language, however the root. Finer distinctions can be made (such as derivations versus borrowings, or direct ancestors versus distant), but these have not been found useful and currently only the general categories exist.

We do not categorize when the etymon is in the same language (e.g. "Category:English terms derived from English" and the like should not exist nor contain any terms). Therefore, either do not use der (or any of the other etymology templates) in these cases or use "-" as the second parameter to suppress categorization. When both languages are the same, the template treats this as a special case automatically and categorises in "twice-borrowed terms":  puts an entry in Category:English twice-borrowed terms. The assumption is that if a language borrowed a term from itself, it must have been borrowed into another language in between.