How to approach love quotes source check: the boomerang

For a love quotes source check, the first question is not who said it. It is what language it was said in, because the answer changes every step that follows.

A very large share of the most-quoted love lines in English arrive through a translator. So the sentence you check was written by someone other than the credited author, in a country and century that may not match the attribution.

Finding the earliest English appearance can then give a confident answer to the wrong question.

What to take away

Three states, three procedures

Written in English. Run the ordinary check: earliest appearance in a dated source. The standard quotation dictionaries, the Oxford Dictionary of Quotations and Bartlett's Familiar Quotations, together with the general search method, settle a line like Elizabeth Barrett Browning's "How do I love thee? Let me count the ways," which is printed in her Sonnets from the Portuguese. Nothing in the rest of this procedure applies.

Translated from another language. Two attributions are now in play, the author's and the translator's, and they are both real, because rendering structure and rendering effect are different tasks with different outputs. The English wording belongs to the translator, and any conclusion you draw from the phrasing is a conclusion about their work.

Claimed as a translation, with no original. More common than it sounds. The line circulates in English with a foreign name attached and has no counterpart in that person's language. The English Rumi is the clearest case: those versions come from Coleman Barks's free renderings of literal translations, collected in books such as The Essential Rumi, so the English wording has no line-for-line counterpart in the Persian. Detecting this state is the main payoff of the whole procedure.

Deciding which of these you have takes one question: does anyone, anywhere, cite a work in the original language? If every reference is English and none names a poem, a collection or a page, treat the third state as your working hypothesis until you can rule it out.

Which translation is this

Assume your wording came from a specific published version, and go and find it.

That outcome, your wording matching no published translation and therefore being a paraphrase of one, is extremely common for short lines, for the reason that short romantic material tends to have no author in the record at all.

Searching the original language

This is the step that settles things, and it is more accessible than people expect.

Take the content words the line needs and look for them in the source language, in that literature's standard vocabulary. Google Books, HathiTrust, the Internet Archive and Wikisource all search inside scanned editions, and every hit carries a publication date you can compare.

Bilingual editions are the most useful tool because they set the original beside a translation, so you find the passage by its English. W. S. Merwin's bilingual edition of Neruda's Twenty Love Poems and a Song of Despair is the standard example.

Catalog search comes next. WorldCat gathers the catalogs of thousands of libraries; the Library of Congress catalog covers the Library's own collections; and a national library in the source country, such as the Bibliothèque nationale de France, the Biblioteca Nacional de España or the Deutsche Nationalbibliothek, records what was actually published there. University collections and digitized literary archives hold the primary texts.

Three practical cautions. Expect some things to have no clean equivalent at all, which is what untranslatability describes, leaving a free rendering where a close one was impossible. Diacritics and non-Latin scripts break naive searching, so try transliterations and unaccented forms too. Older orthography differs from modern spelling, sometimes considerably.

If you do not read the language, a reference librarian or a specialist forum will answer "does this passage exist in this work" far faster than you can guess at it.

What the original-language search returns

The boomerang

A pattern worth knowing because it defeats the obvious test.

A line is invented or heavily reworked in English and attributed to a foreign writer. It becomes popular. Somebody in that writer's own country translates it back into their language, from the English, and publishes it on a page or in a collection. Now a search in the original language returns hits, and the line appears to be confirmed.

Two tells separate a boomerang from an original. Dates: foreign-language appearances are recent and cluster after the English ones, while a real passage appears in editions going back to the work's publication, and the dated hits in Google Books and HathiTrust make that gap visible.

Register: a retranslation reads as modern, with contemporary vocabulary and syntax, sitting oddly against the writer's actual period style, and anyone who reads the language will notice immediately.

The general form of the problem is the same one that arises whenever every apparent source turns out to be citing another source, except that the circular chain here crosses a language boundary, which makes it look like independent confirmation.

Reporting what you found

A translation finding has three separate parts, and running them together is what produces the misleading citations this material is full of.

State all three. A citation that names only the original author, for a sentence composed by a translator a century later, is inaccurate even when the underlying passage is genuine, and love quotes collections that leave the translator out pass the error on to every reader who copies them.

Common questions

Is it wrong to credit the poet rather than the translator?

For the ideas, no. For the wording, yes, because the wording is the translator's work and you are quoting the wording. Naming both costs four words and is simply more accurate.

The translation is old and the translator is not credited anywhere. What then?

Say so: an anonymous nineteenth-century English version is an honest description, and it tells the reader something real about what they are looking at. Sometimes a library catalog will identify the version even when the text you have does not.

A specialist told me the line is not in the original. Can I publish that?

You can publish what you can support: that the line does not appear in the editions you and they checked, and which editions those were. Keep it a statement about what was searched. That is a strong, defensible finding, and it does not require proving a negative about everything the author ever wrote.