The founding poem of English opens on a single word — and translators have shouted it for two hundred years. Lo! What ho! Listen! Now. So. Bro! But the word is literally the question-word what, the famous exclamation mark is an editor's, and the shout may never have been there at all.
In the Old English, the poem begins: Hwæt! We Gardena in geardagum, þeodcyninga, þrym gefrunon — word for word, roughly What! We of the Spear-Danes, in days of yore, of the people-kings, the glory have heard. The very first word of the poem we now call the foundation of English literature is the word hwæt — which is, letter for letter and sense for sense, the ancestor of the modern English what.
And almost no one translates it that way. Open any shelf of Beowulfs and the first word scatters into a small chaos of openings: Lo!, What!, What ho!, Hear me!, Listen!, Yes,, Seamus Heaney's flat Ulster So., Maria Dahvana Headley's Bro! One short syllable, and the carriers cannot agree even on what kind of noise it makes. That disagreement is the subject of this page — and underneath it sits a quieter, stranger claim: that the word may be no shout at all.
Two things travel badly out of the Old English, and it is worth keeping them apart, because they fail independently.
The first is the speech-act. Is hwæt a herald's cry to a noisy hall (Lo!)? A request for attention (Listen!, Hear me!)? A throat-clearing discourse-marker, the Old English of so, anyway (Heaney's So., Kirtlan's Now)? An oath of assent (Yes,, Thorpe's Ay,)? Each is a different opening gesture, and each translator has had to pick one and discard the rest. The word itself does not say.
The second is its grammar, and here is the deep cut. For most of the poem's modern life everyone agreed that hwæt stands outside the sentence — an interjection, flung down, then a full stop, then the poem proper begins: Hwæt! [new clause]. But in 2013 the linguist George Walkden made a careful case that this is a mistake — that hwæt is inside the clause, that Hwæt we ... gefrunon is one long exclamation, and that the line means something closer to How we have heard of the glory of the Spear-Danes! If he is right, the most famous opening punctuation in English — that exclamation mark — is not Old English at all. It is an editor's, added in the nineteenth century. The manuscript never made the shout.
Before the translators, the word. Hwæt is the neuter of hwā, the Old English “who / what” — the same Indo-European root as Latin quod and English what. In surviving Old English it leads three different lives. Tap each. The label where it sits is the tell: two of its lives are firmly inside the sentence; only the third — the “interjection” everyone translates as Lo! — is traditionally placed outside it. That outside placement is exactly what is in dispute.
So the word that opens the poem is, in its other uses, an ordinary interrogative or indefinite pronoun — what. The traditional grammars set the opening hwæt apart as a special fourth thing, the “interjective” or “exclamatory” hwæt: a particle that has shed its meaning and become pure summons. That is the reading two centuries of translators inherited, and shouted.
Here is the heart of it. The line can be punctuated two ways, and the punctuation is not in the manuscript — it is supplied by editors. Move the exclamation mark and watch the sentence change what kind of sentence it is.
The grammar that decides between the two is not a matter of taste — it is a count. Walkden's argument is statistical: he gathered the hwæt-clauses across a corpus (the Old English Bede, Ælfric's Lives of Saints, and the Old Saxon Heliand) and measured where the finite verb falls. If hwæt were an interjection standing outside the clause, the words after it should form an ordinary main clause — and Old English main clauses tend to put the finite verb early, near second position. Instead the verb in hwæt-clauses falls significantly later than in ordinary main clauses (the difference is sharp — Fisher's exact test, p < 0.0001). Beowulf's own opening shows the pattern: the verb gefrunon sits at the very end, “the glory have-heard.” A fresh main clause would not put it there; an exclamative would. And a thing standing outside a clause could not be reaching in to bend its word order — but hwæt does. So it is inside, coloring the whole line as one exclamation, the way German Was does in Was hast du für schöne Augen! — “What beautiful eyes you have!”
For two centuries the loudest word in English literature has been an exclamation mark that no Anglo-Saxon wrote — supplied, the record shows, by an editor in 1833. The shout is editorial.
This is not certain, and the page does not pretend it is — it is one strong, much-discussed argument (it won its field's Hogg Prize the year it appeared), and the older interjection-reading still has defenders. But it reframes everything the translators did. If Walkden is right, the long line of Lo!s and Listen!s did not merely choose the wrong flavor of shout; they heard a shout where the grammar offers a swell.
Now the translations themselves — every one out of copyright, every line transcribed verbatim from a named edition (sources at the foot). For each, the page does two honest things: it highlights the word chosen for hwæt, and it classes that choice by what kind of opening gesture it is — a summons (Lo, Listen), an exclamation (What!, What ho!), a discourse marker (Ay, Now), or the rare degree word that keeps hwæt inside the clause. Sort to watch the fashions move.
The class is a reading, not a measurement — the boundary between a summons and an exclamation is genuinely soft, and the apparatus says so. What is not a reading is the quotation: every word in quotation marks is transcribed exactly from the edition linked beside it.
The verbatim line-up above stops where the public domain reliably does; the famous twentieth- and twenty-first-century versions are in copyright, so they are described, not reproduced — but a translation's opening word is a plain matter of record. Burton Raffel (1963) opened “Hear me!” E. Talbot Donaldson (1966) chose the quiet assent “Yes,” Kevin Crossley-Holland (1968) and, decades later, Roy Liuzza (2000) both settled on the plain imperative “Listen!”; Michael Swanton (1978) reached for the affirming “Indeed”; and J. R. R. Tolkien, whose 1936 lecture made the poem a subject of serious literature and whose own translation appeared in 2014, kept the old “Lo!” And in 2017 the editor Bruce Mitchell did the thing the argument implies: his edition prints Hwæt we with no mark between them at all — the first major edition to un-shout the opening on the page.
Two modern choices are worth naming for how far apart they sit. Seamus Heaney, in the 1999 translation that carried Beowulf back into the wider world, threw out the shout entirely and opened on a single flat syllable he took, he explained in his introduction, from the floor-claiming “So” of his father's Ulster relatives: “So.” — a word that quiets a room and gathers it rather than shouting at it. It is, whether he intended the philology or not, almost exactly the discourse-marker reading: not listen to me but right, then. And Maria Dahvana Headley (2020), translating the poem into the idiom of contemporary masculinity, opened on “Bro!” — which sounds like a provocation and is in fact a precise piece of scholarship: she read hwæt as the bar-side opener of a man about to tell you a story you will not be allowed to leave, the buttonholing so anyway, bro. Heaney and Headley, a generation apart and tonally at war, reached independently for the same insight the grammar had been pointing at — that the first word is less a trumpet than a hand on your shoulder.
For context, the sentence that first word governs. Whether you read it as Hwæt! then a statement, or as one long How-we-have-heard, it runs three lines to its first full stop, naming the inheritance the whole poem will spend itself on: the remembered glory of the Danish kings. The English below is a plain literal crib, not a translation to be admired — and it leaves hwæt in brackets, untranslated, because that single word is the whole question this page asks. Notice the shape it leaves: the object þrym (“glory”) comes before its verb gefrunon (“have heard”) at the line's end — the verb-late shape that is the argument of §II.
Three lines, and the poem has not begun its story — it has only claimed an inheritance: we have heard. The first move of English literature is not an event but a memory of one, and a hush asked for (or a swell announced) so the memory can be told. A translation that mistakes the key of that first word — a parade-ground bark where the poet may have written a gathering-in — has quietly changed the temperature of the whole hall before a single deed is done.
The Old English text. Lines 1–3 are the standard text of Beowulf's opening, given here without editorial macrons (the manuscript marks no vowel length; macrons are a modern teaching aid). The text and a word-by-word gloss are checked against Wikisource and the parallel-text edition at heorot.dk (Slade). Spelling varies by editor (Gardena / Gār-Dena; þeodcyninga / þēod-cyninga); the consonant text is stable.
The word hwæt. That it is the neuter of the interrogative/indefinite pronoun hwā follows the dictionary's headword line for hwā (“hwá, m, f; hwæt; n.”): Bosworth–Toller, An Anglo-Saxon Dictionary, s.v. hwā. The separate entry for hwæt labels it “adv. or interjection” and records the senses used here — interrogative, indefinite (“somewhat, any, at all”), and the interjective “Why, what! ah!”: B–T, s.v. hwæt. The descent ModE what < OE hwæt < PIE *kʷód (neuter of *kʷós), cognate with Latin quod, is the standard etymology. The examples in Instrument I are real Old English — the interrogative is Cædmon's question from the Old English Bede; the indefinite is the standard gif hwæt collocation (B–T sense “somewhat, any, at all”); the interjective is the Beowulf incipit itself. The “interjective” gloss is given as the traditional analysis, which §II then puts in question.
The re-analysis (the record-correction). The argument that hwæt is clause-internal — “not an interjection but an underspecified wh-pronoun introducing an exclamative clause” (the abstract's words), so that the proem reads “How we have heard of the might of the kings…” — is George Walkden, “The status of hwæt in Old English,” English Language and Linguistics 17.3 (2013), 465–488 (doi:10.1017/S1360674313000129; author's PDF); winner of the Richard M. Hogg Prize. Its evidence is statistical: across the Old English Bede, Ælfric's Lives of Saints, and the Old Saxon Heliand, the finite verb falls significantly later in hwæt-clauses than in ordinary main clauses (Fisher's exact test, p < 0.0001) — so hwæt affects the order of the clause after it, and therefore cannot be clause-external. A metrical predecessor is E. G. Stanley, who argued hwæt cannot be a stressed interjection and noted it is never separated from its clause by a manuscript point. It is an argument, much-discussed and not unanimous — the traditional interjection-analysis still has defenders. This page presents Walkden's case as the strongest current reading and labels it as a reading, not a settled fact.
The punctuation. The poem survives in a single manuscript, the Nowell Codex (British Library, Cotton MS Vitellius A.xv), and begins on folio 132r with a large initial H. Old English manuscripts do use the punctus (a point) to break the verse — so the careful claim is not “the manuscript is unpunctuated” but the narrower, sourced one: no point ever divides hwæt from the clause that follows it (Stanley; Walkden). The modern exclamation mark is traceable: per Eric Weiskott, “Making Beowulf Scream: Exclamation and the Punctuation of Old English Poetry,” JEGP 111.1 (2012), 25–41, Thorkelin's 1815 editio princeps used no such mark; the “HWÆT!” convention begins with John Mitchell Kemble's 1833 edition, after which editors freely added exclamation points. (Kemble's separate 1837 translation opened “Lo!” — the edited text's punctuation and the translation's word-choice are distinct facts.)
The translations. Every English line is transcribed verbatim from a public-domain edition, fetched from a primary scan (Internet Archive djvu / Project Gutenberg) linked on each row in Instrument III; the word marked as rendering hwæt is the translator's own. Two were quoted from a later printing of an unchanged earlier translation (Garnett's 4th-ed. scan of the 1882 text; Morris & Wyatt's 1898 printing of the 1895 Tale); both are noted. The orthodoxy was conscious: John Earle, choosing “What ho!” in 1892, appended a note that “this what is not interrogative, but interjectional” — the interjection-reading stated outright, the very reading Walkden's count would later contest. The classification (summons / exclamation / discourse-marker / degree) is the page's reading and is deliberately coarse; the boundaries are soft and are flagged as such — Ay and Now are grouped as discourse markers (assent/transition), not shouts.
The honest edges. “Class” is an interpretive label, not a measurement — Lo! and What! shade into each other, and a translator's tone can pull a word across the line. The claims about modern (in-copyright) translators are limited to their opening word, a matter of record; their texts are not reproduced. The reading of Heaney's “So.” and Headley's “Bro!” as discourse-markers follows the translators' own published notes. And the Walkden re-analysis, as above, is offered as argument, not theorem.
Type. Old English and Latin both set in Fraunces and Martian Mono (SIL Open Font License), self-hosted — Old English needs no special face, only the letters þ ð æ that the Latin font already carries. No third-party requests; everything is served first-party.
An entry in the Translation-Criticism Venue, and the deliberate companion to The First Word Is Rage — the first word of the other founding epic of the West. There, Homer's μῆνιν: a word whose register and position English cannot keep. Here, hwæt: a word whose very part of speech — shout or swell, outside the sentence or in — English cannot decide. Its other companions: The Way That Can Be Told (Laozi's first line, and an editorial artifact made a toggle), The Eye of the Needle (where the distortion lives not in the translation but in the apparatus around it), and The River That Stays.
← back to the ground