How to use an etymological dictionary

Etymology (the history of words) fascinates many people. The question where a word comes from, what it used to mean, and when it is first attested in writing is interesting not only for historical linguists. As a historical linguist who also does sci-comm, I often get asked about the etymology of a specific word. I noticed that non-linguists usually face two problems when it comes to etymologies: They do not know where to find them or they are unsure how to interpret etymological resources. This blog post is a guide on how to use etymological dictionaries and other resources that inform about etymologies.

Where can I find etymologies?

When you ask a linguist about the etymology of a word (let’s say turban), chances are low that they will establish the etymology themself. Instead, they will likely fall back on work by other linguists who compiled resources on etymologies. These usually come in the form of etymological dictionaries, or nowadays more often online databases. For most major and many smaller languages etymological resources already exist. Many of them can be freely accessed online, also by non-linguists. So there is no reason why you cannot look up an etymology yourself, as long as you follow some basic rules.

The question how etymologies are established goes beyond this blog post. One problem with etymological resources is that their content is very difficult to evaluate if you are not an expert on the languages discussed. This is also true for linguists. Since German studies is not my field of research, I am not able to check whether a suggested etymology for a German word is actually correct. That would require me to do my own research on historical sources and look up (or even sketch myself) the history of the language. All I can do is assess the overall plausibility of an etymology and trust that the author of the resource adhered to scientific standards.

If, for example, the author claims that turban comes from English towel band than I would question that based on my knowledge of the history of turbans (which did not originate in Europe) and of phonological history. How does the author explain the /r/ that occurs in turban but not in towel band?, etc. If they provide no explanation that can be a sign that this is not a well established etymology. If possible, check several resources and then point out where they agree or differ from each other.

Let’s look at some examples for etymological resources.

One very accessible resource is wiktionary. Wiktionary contains data for hundreds of languages and since entries are linked you can click your way through the history of a word. There are two big problems with wiktionary: Firstly, sources are usually not declared and, secondly, quality differs from language to language.

Another free online etymological resource for English is the Online Etymological Dictionary. It is compiled by Douglas R. Harper and based on several etymological dictionaries. The website usually provides more information than wiktionary and every entry includes the google ngram. That is a graph that shows how frequently a word has been used over the years.

The most widely regarded etymological resource for English is the Oxford English Dictionary (OED). Every entry in the OED includes an etymology section. The main problem here is that it is not freely accessible, you need a subscription (or the actual paper volume).

The first edition of the OED is now in public domain and you can download it for free. But notice that etymologies can become outdated and therefore they should be checked with more recent proposals.

These are just three examples of etymological resources for English. You can check out this list for more resources for multiple languages.

How do I use an etymological resource?

The moment has come: you are sitting in front of an etymological resource ready to look up an etymology. You may find that the entry you are looking at is not self-explanatory. Often there are abbreviations or technical terms. Fortunately, most etymological resources use a similar style. In this section, I will explain how to interpret traditional etymological dictionaries, online resources that use a similar format as traditional dictionaries, and modern databases.

Traditional etymological dictionaries

Let’s have a look at the first edition of the Oxford Etymological Dictionary. This is the entry for turban.

The word the entry is about is called a headword. It is followed by the phonetic form (the pronunciation) and the word class. After that, all attested forms in English are listed.

The etymology section reiterates the history of the word turban, starting out in Persian and then making its way via Turkish into several European languages before reaching English. Importantly, forms are given in italics.

One obstacle here are the abbreviations such as OIt (= Old Italian) and obs. F (= obsolete French). A list of abbreviations is found at the beginning of the dictionary. Another difficulty is to distinguish languages from which English turban descends from languages that also borrowed the word. The forms after the last semicolon in the first sentence (”early mod. Du. turbant…”) are such borrowings into other languages. English did not borrowed turban from Dutch or German, the forms are listed for comparative purposes. Following the “path” of the word are further information, in this case, about the change from /l/ to /r/.

Also interesting are the attestation at the bottom of the entry. They are ordered chronologically.

Traditional comparative dictionaries

The Oxford English Dictionary (OED) offer etymologies for a single language. Other sources compare multiple languages; these are called comparative dictionaries. The goal of these dictionaries is to identify words in related languages that originate from the source word (etymon). Therefore, comparative dictionaries are typically organized by etymons.

This is the entry for *patér ‘father’ from the Vergleichendes Wörterbucher der Indogermanischen Sprachen by August Fick, published in 1890 (p. 469). The headword is the reconstruction, followed by the gloss and inflectional forms (locative, dative and accusative plural). Then follow attested forms in different languages such as Greek, Latin and Old Irish.

Online etymological resources

Online etymological resources such as the Online Etymological Dictionary resemble traditional dictionaries in that they contain entries for individual words. The entry for turban is very similar to the entry in the OED.

The title of the entry contains the headword and the word class. The entry starts with the period in which the word is first attested. Then follows the etymology and information about the /l/ to /r/ change. The entry closes with the definition and a note on related entries. Below the entry is a link to a list of word first attested around the same time.

The entry in wiktionary starts with alternative forms (similar to the attested forms section in the OED). Then follows the etymology section which reiterates the path of the word through different languages. It also links to the related word tulip. After the etymology follows the pronunciation and the definition.

The online version of the OED has an extra section for etymology (only accessible with a subscription). It is still very similar to the entry in the first edition. The section is headed by a summary of the entry. The actual eymology is followed by a note on the first attestation of the word. Then follow alternative forms in English and a reference to tulip.

Modern databases

Modern etymological resources do not always come in the form of entries mimicking traditional dictionaries. One example is the Austronesian Comparative Dictionary Online. It is structured along cognate sets (sets of words belonging to the same etymon, similar to the traditional comparative dictionary). You can for example search for a gloss (a meaning) and then you will be showed all Austronesian reconstructions for this gloss.

Each reconstruction has its own page that contains all attested descendants of the reconstruction. Sometimes, there are also notes on the set displayed in the sidebar. The attested forms can be found below the map.

Comparing etymological resources

Comparing several etymological resources is facilitated by websites that provide several resources. One example is etymologiebank.nl for Dutch. It displays entries from several sources for the same modern Dutch word alongside each other. In the case of tulband ‘turban’ you can compare five etymological resources, as well as dialect dictionaries and more. Unfortunately, this website is pretty unique. I actually do not know of any similar websites for other languages.

More things to keep in mind

I mentioned above that distinguishing between sources of borrowing and comperanda (words for comparison) can be difficult. Sometimes people believe that whenever a form for an ancient language is listed in an entry, the respective word in English (or whatever language) must be borrowed from this “older” language. Take a look at the entry for three in the Online Etymological Dictionary. It also lists forms from Sanskrit, Greek and Latin. However, English did not borrow the word three from one of these languages. The crucial phrase here is “source also of”. The forms are listed in the entry because they are derived from the same Proto-Indo European root as the English word.

Sometimes, not every word has an entry in an etymological dictionary. Sometimes, for example nouns derived from verbs are discussed in the entry of the respective verb. The Etymologisches Wörterbuch des Deutschen is one example. The noun Stiftung is discussed in the entry for the verb stiften.

The limits of etymology

In this last section I would like to point out some caveats about the limits of etymologies. Even when using a trustworthy source and interpreting it in the proper way, people may still misuse an etymology.

One common misunderstanding about etymologies is that they can reveal something about the “true” meaning of a word. This is known as the etymological fallacy . In fact, the history of a word sometimes tells us nothing about its current meaning or usage. For example, when first hearing the word butterfly, some people may assume that it is a fly somehow connected to butter. This is a rather harmless example of an etymological fallacy but sometimes people insist that you cannot use a word in a certain sense because that is not what it used to mean. For example, someone being angry about people using transpire to mean ‘to happen’ because in Latin, whence it comes from, transpirare means ‘to breathe across’.

In other cases, etymology can help us understand a word better. For example, the word gospel is derived from Old English god + spell. Spell used to mean ‘speech’ and god is ‘good, hence godspell = ‘good news’. But again, this does not provide information on the current meaning of the word, it only brings light to its history.

Another misunderstanding concerns the handling of different proposed etymologies for the same word. If there are several proposed etymologies, you cannot simply choose the one you like best. I often see people treating multiple etymologies as if they were of equal value.With regard to our turban example, I can imagine someone posting on twitter along the following lines: “Turban has been argued to come from Persian dulband but other people believe it is from English towel band.” Etymologies are not opinions. If there is no historical evidence to support them, they are not well established, even if they sound cuter.

---

---