This article needs additional citations for verification. (October 2011) (Learn how and when to remove this template message)
|Part of a series on|
The grammar of Old English is quite different from that of Modern English, predominantly by being much more inflected. As an old Germanic language, Old English has a morphological system that is similar to that of the hypothetical Proto-Germanic reconstruction, retaining many of the inflections thought to have been common in Proto-Indo-European and also including characteristically Germanic constructions such as the umlaut.
Among living languages, Old English morphology most closely resembles that of modern Icelandic, which is among the most conservative of the Germanic languages; to a lesser extent, the Old English inflectional system is similar to that of modern German.
Nouns, pronouns, adjectives and determiners were fully inflected with five grammatical cases (nominative, accusative, genitive, dative, and instrumental), two grammatical numbers (singular and plural) and three grammatical genders (masculine, feminine, and neuter). First- and second-person personal pronouns also had dual forms for referring to groups of two people, in addition to the usual singular and plural forms. The instrumental case was somewhat rare and occurred only in the masculine and neuter singular. It was often replaced by the dative. Adjectives, pronouns and (sometimes) participles agreed with their antecedent nouns in case, number and gender. Finite verbs agreed with their subject in person and number.
Nouns came in numerous declensions (with many parallels in Latin, Ancient Greek and Sanskrit). Verbs came in nine main conjugations (seven strong and two weak), all with numerous subtypes, as well as a few additional smaller conjugations and a handful of irregular verbs. The main difference from other ancient Indo-European languages, such as Latin, is that verbs could be conjugated in only two tenses (vs. the six "tenses," really tense/aspect combinations, of Latin), and they have no synthetic passive voice although it still existed in Gothic.
The grammatical gender of a given noun does not necessarily correspond to its natural gender, even for nouns referring to people. For example, sēo sunne (the Sun) was feminine, se mōna (the Moon) was masculine, and þæt ƿīf "the woman/wife" was neuter. (Compare modern German die Sonne, der Mond, das Weib.) Pronominal usage could reflect either natural or grammatical gender, when it conflicted.
Strong verbs use the Germanic form of conjugation known as ablaut. In this form of conjugation, the stem of the word changes to indicate the tense. Verbs like this persist in modern English; for example sing, sang, sung is a strong verb, as are swim, swam, swum and choose, chose, chosen. The root portion of the word changes rather than its ending. In Old English, there were seven major classes of strong verb; each class has its own pattern of stem changes. Learning these is often a challenge for students of the language, though English speakers may see connections between the old verb classes and their modern forms.
The classes had the following distinguishing features to their infinitive stems:
|Class||Root weight||Infinitive||First preterite||Second preterite||Past participle|
|II||ēo or ū||ēa||u||o|
|III||see table below|
|VII||heavy||ō, ā, ēa, a (+nC), ea (+rC/lC), occ. ǣ||ē or ēo||same as infinitive|
The first preterite stem is used in the preterite, for the first- and third-person singular. The second preterite stem is used for second-person singular, and all persons in the plural (as well as the preterite subjunctive). Strong verbs also exhibit i-mutation of the stem in the second- and third-person singular in the present tense.
The third class went through so many sound changes that it was barely recognisable as a single class. The first was a process called 'breaking'. Before ⟨h⟩, and ⟨r⟩ + another consonant, ⟨æ⟩ turned into ⟨ea⟩, and ⟨e⟩ to ⟨eo⟩. Also, before ⟨l⟩ + another consonant, the same happened to ⟨æ⟩, but ⟨e⟩ remained unchanged (except before combination ⟨lh⟩).
The second sound change to affect it was the influence of palatal sounds ⟨g⟩, ⟨c⟩, and ⟨sc⟩. These turned preceding ⟨e⟩ and ⟨æ⟩ to ⟨ie⟩ and ⟨ea⟩, respectively.
The third sound change turned ⟨e⟩ to ⟨i⟩, ⟨æ⟩ to ⟨a⟩, and ⟨o⟩ to ⟨u⟩ before nasals.
Altogether, this split the third class into five sub-classes:
|Sub-class||Infinitive||First preterite||Second preterite||Past participle|
Regular strong verbs were all conjugated roughly the same, with the main differences being in the stem vowel. Thus stelan "to steal" represents the strong verb conjugation paradigm.
Weak verbs are formed by adding alveolar (t or d) endings to the stem for the past and past-participle tenses. Examples include love, loved and look, looked.
Originally, the weak ending was used to form the preterite of informal, noun-derived verbs such as often emerge in conversation and which have no established system of stem-change. By nature, these verbs were almost always transitive, and even today, most weak verbs are transitive verbs formed in the same way. However, as English came into contact with non-Germanic languages, it invariably borrowed useful verbs which lacked established stem-change patterns. Rather than inventing and standardizing new classes or learning foreign conjugations, English speakers simply applied the weak ending to the foreign bases.
The linguistic trends of borrowing foreign verbs and verbalizing nouns have greatly increased the number of weak verbs over the last 1,200 years. Some verbs that were originally strong (for example help, holp, holpen) have become weak by analogy; most foreign verbs are adopted as weak verbs; and when verbs are made from nouns (for example "to scroll" or "to water") the resulting verb is weak. Additionally, conjugation of weak verbs is easier to teach, since there are fewer classes of variation. In combination, these factors have drastically increased the number of weak verbs, so that in modern English weak verbs are the most numerous and productive form, although occasionally a weak verb may turn into a strong verb through the process of analogy, such as sneak (originally only a noun), where snuck is an analogical formation rather than a survival from Old English.
There are three major classes of weak verbs in Old English. The first class displays i-mutation in the root, and the second class none. There is also a third class explained below.
Class-one verbs with short roots exhibit gemination of the final stem consonant in certain forms. With verbs in ⟨r⟩, this appears as ⟨ri⟩ or ⟨rg⟩, where ⟨i⟩ and ⟨g⟩ are pronounced [j]. Geminated ⟨f⟩ appears as ⟨bb⟩, and that of ⟨g⟩ appears as ⟨cg⟩. Class-one verbs may receive an epenthetic vowel before endings beginning in a consonant.
Where class-one verbs have gemination, class-two verbs have ⟨i⟩ or ⟨ig⟩, which is a separate syllable pronounced [i]. All class-two verbs have an epenthetic vowel, which appears as ⟨a⟩ or ⟨o⟩.
In the following table, three verbs are conjugated. Sƿebban "to put to sleep" is a class-one verb exhibiting gemination and an epenthetic vowel. Hǣlan "to heal" is a class-one verb exhibiting neither gemination nor an epenthetic vowel. Sīðian "to journey" is a class-two verb.
|Tense/mood||Pronoun||'put to sleep'||'heal'||'journey'|
|tō sƿebbanne||tō hǣlanne||tō sīðianne|
During the Old English period, the third class was significantly reduced; only four verbs belonged to this group: habban 'have', libban 'live', secgan 'say', and hycgan 'think'. Each of these verbs is distinctly irregular, though they share some commonalities.
|tō hæbbenne||tō libbenne||tō secgenne||tō hycgenne|
|þū||hæfst, hafast||lifast, leofast||segst, sagast||hygst, hogast|
|hē/hit/hēo||hæfð, hafað||lifað, leofað||segð, sagað||hyg(e)d, hogað|
|Past indicative||ic/hē/hit/hēo||hæfde||lifde, leofode||sægde||hog(o)de, hygde|
|þū||hæfdest||lifdest, leofodest||sægdest||hog(o)dest, hygdest|
|Plural||hæfdon||lifdon, leofodon||sægdon||hog(o)don, hygdon|
|Present subjunctive||Singular||hæbbe||libbe, lifge||secge||hycge|
|Past subjunctive||Singular||hæfde||lifde, leofode||sægde||hog(o)de, hygde|
|Plural||hæfden||lifde, leofoden||sægden||hog(o)den, hygden|
|Imperative||Singular||hafa||leofa||sæge, saga||hyge, hoga|
|Present participle||hæbbende||libbende, lifgende||secgende||hycgende|
The preterite-present verbs are a class of verbs which have a present tense in the form of a strong preterite and a past tense like the past of a weak verb. These verbs derive from the subjunctive or optative use of preterite forms to refer to present or future time. For example, ƿitan, "to know" comes from a verb which originally meant "to have seen" (cf. OE ƿise "manner, mode, appearance"; Latin videre "to see" from the same root). The present singular is formed from the original singular preterite stem and the present plural from the original plural preterite stem. As a result of this history, the first-person singular and third-person singular are the same in the present.
Few preterite-present verbs appear in the Old English corpus, and the forms marked with an asterisk are unattested reconstructions, formed by analogy.
In spite of heavy irregularities, there are four groups of similarly-conjugated verbs:
Note that the Old English meanings of many of the verbs are significantly different from that of the modern descendants; in fact, the verbs "can, may, must", and to a lesser extent "thurf, durr" appear to have chain shifted in meaning.
|Conjugation||Pronoun||'know, know how to'||'be able to, can'||'be obliged to, must'||'know'||'own'||'avail'||'dare'||'remember'||'need'||'be allowed to, may'||'grant, allow, wish'||'have use of, enjoy'|
|Modern descendant||can, could||may, might||shall, should||wit, wost (archaic)||owe, *aught||dow, *dight (archaic)||*dure, dare||*(i-)mune[clarification needed]||thurf, tharf (archaic)||mote (archaic), must||*ann, ould||*(i-)now, (i-)night[clarification needed]|
|tō cunnenne||tō magenne||tō sculenne||tō ƿitenne||tō āgenne||tō dugenne||tō durrenne||tō ge-munenne||tō þurfenne||tō mōtenne||tō unnenne||tō ge-/benugenne|
|Past indicative||ic||cūðe||mihte, meahte||sceolde||ƿisse, ƿiste||āhte||dohte||dorste||gemunde||þorfte||mōste||uðe||benohte|
|Past subjunctive||Singular||cūðe||mihte||sceolde||ƿisse, ƿiste||*āhte||*dohte||*dorste||gemunde||*þorfte||*mōste||*uðe||*benohte|
Additionally, there is a further group of four verbs which are anomalous: "want" (modern "will"), "do", "go" and "be". These four have their own conjugation schemes which differ significantly from all the other classes of verb. This is not especially unusual: "want", "do", "go", and "be" are the most commonly used verbs in the language, and are very important to the meaning of the sentences in which they are used. Idiosyncratic patterns of inflection are much more common with important items of vocabulary than with rarely used ones.
Dōn 'to do' and gān 'to go' are conjugated alike; ƿillan 'to want' is similar outside of the present tense.
|tō dōnne||tō gānne||tō willanne|
The verb 'to be' is actually composed of three different stems:
The present forms of ƿesan are almost never used. Therefore, ƿesan is used as the past, imperative, and present participle versions of sindon, and does not have a separate meaning. The bēon forms are usually used in reference to future actions. Only the present forms of bēon contrast with the present forms of sindon/ƿesan in that bēon tends to be used to refer to eternal or permanent truths, while sindon/ƿesan is used more commonly to refer to temporary or subjective facts. This semantic distinction (made only during the present tense) was lost as Old English developed into modern English, so that the modern verb 'to be' is a single verb which takes its present indicative forms from sindon, its past indicative forms from ƿesan, its present subjunctive forms from bēon, its past subjunctive forms from ƿesan, and its imperative and participle forms from bēon. (Modern German had an analogous, but even more complicated, development for its verb sein.) In late OE and ME, the form earon/earun, from the Old Norse erun, replaced bēoþ and sind (See also List of English words of Old Norse origin).
Old English is an inflected language, and as such its nouns, pronouns, adjectives and determiners must be declined in order to serve a grammatical function. A set of declined forms of the same word pattern is called a declension. As in several other ancient Germanic languages, there are five major cases: nominative, accusative, dative, genitive and instrumental.
The small body of evidence available for Runic texts suggests that there may also have been a separate locative case in early or Northumbrian forms of the language (e.g., ᚩᚾ ᚱᚩᛞᛁ on rodi "on the Cross").
In addition to inflection for case, nouns take different endings depending on whether the noun was in the singular (for example, hring "one ring"') or plural (for example, hringas "many rings"). Also, some nouns pluralize by way of Umlaut, and some undergo no pluralizing change in certain cases.
Nouns are also categorized by grammatical gender – masculine, feminine, or neuter. In general, masculine and neuter words share their endings, while feminine words have their own subset of endings. The plural of some declension types distinguishes between genders, e.g., a-stem masculine nominative plural stanas "stones" vs. neuter nominative plural scipu "ships" and word "words"; or i-stem masculine nominative plural sige(as) "victories" vs. neuter nominative plural sifu "sieves" and hilt "hilts".
Furthermore, Old English nouns are divided as either strong or weak. Weak nouns have their own endings. In general, weak nouns are less complex than strong nouns, since they had begun to lose their system of declension. However, the various noun classes are not totally distinct from one another, and there is a great deal of overlap between them.
Descriptions of Old English language grammars often follow the NOM-ACC-GEN-DAT-INST case order.
Here are the strong declensional endings and examples for each gender:
For the '-u/–' forms above, the '-u' is used with a root consisting of a single short syllable or ending in a long syllable followed by a short syllable, while roots ending in a long syllable or two short syllables are not inflected. (A long syllable contains a long vowel or is followed by two consonants. Note also that there are some exceptions; for example, feminine nouns ending in -þu such as strengþu 'strength'.)
Note the syncope of the second e in engel when an ending follows. This syncope of the vowel in the second syllable occurs with two-syllable strong nouns, which have a long vowel in the first syllable and a second syllable consisting of a short vowel and single consonant (for example, engel, ƿuldor 'glory', and hēafod 'head'). However, this syncope is not always present, so forms such as engelas may be seen.
Here are the weak declensional endings and examples for each gender:
In addition, masculine and neuter nouns whose main vowel is short æ and end with a single consonant change the vowel to a in the plural (a result of the phonological phenomenon known as Anglo-Frisian brightening):
Some masculine and neuter nouns end in -e in their base form. These drop the -e and add normal endings. Note that neuter nouns in -e always have -u in the plural, even with a long vowel:
Nouns ending in -h lose this when an ending is added, and lengthen the vowel in compensation (this can result in compression of the ending as well):
Nouns whose stem ends in -ƿ change this to -u or drop it in the nominative singular. (Note that this '-u/–' distinction depends on syllable weight, as for strong nouns, above.)
|Accusative||sinƿe||sinƿa, -e||lǣsƿe||lǣsƿa, -e|
A few nouns follow the -u declension, with an entirely different set of endings. The following examples are both masculine, although feminines also exist, with the same endings (for example duru 'door' and hand 'hand'). Note that the '-u/–' distinction in the singular depends on syllable weight, as for strong nouns, above.
There are also some nouns of the consonant declension, which show i-umlaut in some forms.
|Genitive||fōtes||fōta||hnyte, hnute||hnuta||bēc, bōce||bōca|
|Dative||fēt, fōte||fōtum||hnutum||bēc, bōc||bōcum|
Other such nouns include (with singular and plural nominative forms given):
Masculine: tōþ, tēþ 'tooth'; mann, menn 'man'; frēond, frīend 'friend'; fēond, fīend 'enemy' (cf. 'fiend')
Feminine: studu, styde 'post' (cf. 'stud'); hnitu, hnite 'nit'; āc, ǣc 'oak'; gāt, gǣt 'goat'; brōc, brēc 'leg covering' (cf. 'breeches'); gōs, gēs 'goose'; burg, byrg 'city' (cf. 'borough', '-bury' and German cities in -burg); dung, dyng 'prison' (cf. 'dungeon' by way of French and Frankish); turf, tyrf 'turf'; grūt, grȳt 'meal' (cf. 'grout'); lūs, lȳs 'louse'; mūs, mȳs 'mouse'; neaht, niht 'night' Feminine with loss of -h in some forms: furh, fyrh 'furrow' or 'fir'; sulh, sylh 'plough'; þrūh, þrȳh 'trough'; ƿlōh, ƿlēh 'fringe'. Feminine with compression of endings: cū, cȳ 'cow' (cf. dialectal plural 'kine')
Neuter: In addition, scrūd 'clothing, garment' has the umlauted dative-singular form scrȳd.
|Nominative-Accusative||fæder||fæd(e)ras||brōðor||(ge)brōðor||mōdor||mōdra/mōdru||sƿeostor||(ge)sƿeostor, -tru, -tra||dohtor|
Other such nouns: ǣg, ǣgru egg (ancestor of the archaic or dialectal form ey, plural eyren; the form egg is a borrowing from Old Norse); bread, breadru 'crumb'; cealf, cealfru 'calf'; cild 'child' has either the normal plural cild or cildru (cf. 'children', with -en from the weak nouns); hǣmed, hǣmedru 'cohabitation'; speld, speldru 'torch'.
Adjectives in Old English are declined using the same categories as nouns: five cases (nominative, accusative, genitive, dative, and instrumental), three genders (masculine, feminine, neuter), and two numbers (singular, plural). In addition, they can be declined either strong or weak. The weak forms are used in the presence of a definite or possessive determiner, while the strong ones are used in other situations. The weak forms are identical to those for nouns, while the strong forms use a combination of noun and pronoun endings:
For the '-u/–' forms above, the distinction is the same as for strong nouns.
Example of the strong adjective declension: gōd 'good'
Note that the same variants described above for nouns also exist for adjectives. The following example shows both the æ/a variation and the -u forms in the feminine singular and neuter plural:
The following shows an example of an adjective ending with -h:
The following shows an example of an adjective ending with -ƿ:
Old English had two main determiners: se, which could function as both 'the' or 'that', and þes for 'this'.
Modern English 'that' descends from the neuter nominative/accusative form, and 'the' from the masculine nominative form, with 's' replaced analogously by the 'th' of the other forms. The feminine nominative form was possibly the source of Modern English 'she'.
Most pronouns are declined by number, case and gender; in the plural form most pronouns have only one form for all genders. Additionally, Old English pronouns preserve the dual form (which is specifically for talking about groups of two things, for example "we two" or "you two" or "they two"). These were uncommon even then, but remained in use throughout the period.
|Accusative||mec, mē||ūsic, ūs||uncit, unc|
|Accusative||þec, þē||ēoƿic, ēoƿ||incit, inc|
Many of the forms above bear strong resemblances to their contemporary English language equivalents: for instance in the genitive case ēoƿer became "your", ūre became "our", mīn became "mine". Some forms do not match their modern equivalents due to dissimilation. The feminine nominative hēo was at some point replaced with the feminine nominative article sēo, yielding "she"; whereas the h in plural forms such as hīe was replaced with þ under Norse influence as it evolved (a slower development that was not complete until well into the Middle English period), yielding "they, them, their".
Prepositions (like Modern English words by, for, and with) often follow the word which they govern, in which case they are called postpositions. Also, if the object of a preposition was marked in the dative case, a preposition may conceivably be located anywhere in the sentence.
The following is a list of prepositions in the Old English language. Many of them, particularly those marked "etc.", are found in other variant spellings. Prepositions may govern the accusative, genitive, dative or instrumental cases.
|æfter||after; along, through, during; according to, by means of; about||Ancestor of modern after; related to Dutch achter = behind, after|
|ǣr||before||Related to modern German eher and Dutch (vooral)eer, ancestor of modern ere|
|æt||at, to, before, next, with, in, for, against; unto, as far as||Ancestor of modern at|
|and||against, before, on||Related to Dutch aan (e.g. tegenaan)|
|andlang||along||Ancestor of modern along, related to modern German entlang|
|beæftan||after, behind; without||Ancestor of modern (nautical) abaft|
|be, bī||by, near to, to, at, in, on, upon, about, with; of, from, about, touching, concerning; for, because of, after, by, through, according to; beside, out of||Related to modern German bei, ancestor of modern by|
|befōran||before||Ancestor of modern before, related to modern German bevor|
|begeondan||beyond||Ancestor of modern beyond|
|behindan||behind||Ancestor of modern behind, related to modern German hinter|
|beinnan||in, within||Related to modern German and Dutch binnen|
|beneoðan||beneath||Ancestor of modern beneath, cf. Dutch beneden|
|betƿeonum, betƿeox, etc.||betwixt, between, among, amid, in the midst.||Ancestors of modern between and betwixt respectively|
|būfan||above||Ancestor of modern above through compound form onbúfan|
|būtan||out of, against; without, except||Related to modern Dutch buiten|
|eāc||with, in addition to, besides||Related to modern German auch and Dutch ook, ancestor of modern (archaic) eke|
|for||for, on account of, because of, with, by; according to; instead of.||Ancestor of modern for, related to modern German für|
|fōr, fōre||before||Related to modern German vor, bevor and Dutch voor|
|fram||from; concerning, about, of||Ancestor of modern from|
|gemang||among||Ancestor of modern among|
|geond||through, throughout, over, as far as, among, in, after, beyond||Ancestor of modern yonder through comparative form geondra. Related to Dutch ginds and (archaic) ginder|
|in||in, on; into, to||Ancestor of modern in, related to German and Latin in|
|innan||in, into, within, from within||Related to modern German innen|
|intō||into||Ancestor of modern into|
|mid||with, against||Ancestor of modern amid through related form onmiddan (cf. Dutch onmiddellijk), related to modern German mit|
|neāh||near||Ancestor of modern nigh. Dutch naar (via nader), German nah(e)|
|of||of, from, out of, off||Ancestor of modern of and off|
|ofer||above, over; upon, on; throughout; beyond, more than||Ancestor of modern over|
|on||on; in, at||Ancestor of modern on|
|onbūtan||about||Ancestor of modern about|
|ongeagn, etc.||opposite, against; towards; in reply to||Ancestor of modern again. Related to German entgegen|
|oþ||to, unto, up to, as far as||Related to Dutch and Frisian op and to modern German auf|
|samod||with, at||Related to modern German samt, mitsamt|
|tō||to, at||Ancestor of modern to, related to modern German zu|
|tōeācan||in addition to, besides|
|tōforan||before||Related to Dutch tevoren, German zuvor|
|tōgeagnes||towards, against||Related to Dutch tegen|
|tōmiddes||in the midst of, amidst||Related to Dutch temidden|
|tōƿeard||toward||Ancestor of modern toward|
|þurh||through||Related to modern German durch, ancestor of modern through|
|under||under||Ancestor of modern under, related to modern German unter|
|underneoþan||underneath||Ancestor of modern underneath|
|uppan||upon, on||Not the ancestor of modern upon, which came from "up on".|
|ūtan||without, outside of||Related to modern Swedish utan, Dutch uit and German außen, außer. The adverbial form ūt is the ancestor of modern out.|
|ƿið||towards, to; with, against; opposite to; by, near||Ancestor of modern with|
|ƿiðer||against||Related to modern German wider|
|ƿiðinnan||within||Ancestor of modern within|
|ƿiðūtan||without, outside of||Ancestor of modern without|
|ymb, ymbe||about, by||Related to modern German um and Latin ambi|
|ymbūtan||about, around; concerning|
Old English syntax was similar in many ways to that of Modern English. However, there were some important differences. Some were simply consequences of the greater level of nominal and verbal inflection, and word order was generally freer. There are also differences in the default word order and in the construction of negation, questions, relative clauses and subordinate clauses.
There was some flexibility in word order of Old English since the heavily inflected nature of nouns, adjectives, and verbs often indicated the relationships between clause arguments. Scrambling of constituents was common. Even sometimes scrambling within a constituent occurred, as in Beoƿulf line 708 ƿrāþum on andan:
|hostile (Dative Singular)||on/with||malice (Dative Singular)|
|"with hostile malice"|
Something similar occurs in line 713 in sele þām hēan "in the high hall" (lit. "in hall the high").
Note how the words ond Ƿestseaxna ƿiotan "and the West Saxon counselors" (lit. "and (the) counselors of (the) West Saxons") have been extraposed from (moved out of) the compound subject they belong in, in a way that would be impossible in modern English. In Old English, case inflection preserves the meaning: the verb beniman "to deprive" (appearing in this sentence in the form benam, "[he] deprived") needs a word in the genitive case to show what someone or something is deprived of, which in this sentence is rīces "of kingdom" (nominative rīce, "kingdom"), whereas ƿiotan "counselors" is in the nominative case and therefore serves a different role entirely (the genitive of it would be ƿiotana, "of counselors"); for this reason the interpretation that Cyneƿulf deprived Sigebryht of the West Saxon counselors was not possible for speakers of Old English. Note that the Old English sentence still isn't in theory perfectly unambiguous, as it contains one more word in the genitive: Ƿestseaxna ("of West Saxons", nominative Ƿestseaxan "West Saxons"), and the form ƿiotan "counselors" may also represent the accusative case in addition to the nominative, thus for example creating the grammatical possibility of the interpretation that Cyneƿulf also took the West Saxons away from the counselors, but this would have been difficult to conceive.
Main clauses in Old English tend to have a verb-second (V2) order, where the verb is the second constituent in a sentence, regardless of what comes first. There are echoes of this in modern English: "Hardly did he arrive when ...", "Never can it be said that ...", "Over went the boat", "Ever onward marched the weary soldiers ...", "Then came a loud sound from the sky above". In Old English, however, it was much more extensive, like the word order in modern German. If the subject appears first, there is an SVO order, but it can also yield orders such as OVS and others. In questions VSO was common, see below.
In subordinate clauses, however, the word order is completely different, with verb-final constructions the norm, again as in German. Furthermore, in poetry, all the rules were frequently broken. In Beoƿulf, for example, main clauses frequently have verb-initial or verb-final order, and subordinate clauses often have verb-second order. (However, in clauses introduced by þā, which can mean either "when" or "then", and where word order is crucial for telling the difference, the normal word order is nearly always followed.)
Those linguists who work within the Chomskyan transformational grammar paradigm often believe that it is more accurate to describe Old English (and other Germanic languages with the same word-order patterns like modern German) as having underlying subject-object-verb (SOV) ordering. According to this theory, all sentences are initially generated using this order, but in main clauses, the verb is moved back to the V2 position (technically, the verb undergoes V-to-T raising). That is said to explain the fact that Old English allows inversion of subject and verb as a general strategy for forming questions, while modern English uses this strategy almost only with auxiliary verbs and the main verb "to be", requiring do-support in other cases.
Because of its similarity with Old Norse, it is believed that most of the time the word order of Old English changed when asking a question, from SVO to VSO. While many purport that Old English had free word order, this is not quite true, as there were conventions for the positioning of subject, object and verb in clause.
Old English did not use forms equivalent to "who, when, where" in relative clauses (as in "The man whom I saw") or subordinate clauses ("When I got home, I went to sleep").
Instead, relative clauses used one of the following:
Preposition-fronting ("The man with whom I spoke") did not normally occur.
Subordinate clauses tended to use correlative conjunctions, e.g.
The word order usually distinguished the subordinate clause (with verb-final order) from the main clause (with verb-second word order).
Besides þā ... þā ..., other correlative conjunctions occurred, often in pairs of identical words, e.g.:
The phonology of Old English is necessarily somewhat speculative, since it is preserved purely as a written language. Nevertheless, there is a very large corpus of Old English, and the written language apparently indicates phonological alternations quite faithfully, so it is not difficult to draw certain conclusions about the nature of Old English phonology.