Weak verbs form a category separate from strong verbs. Whereas vowel gradation (ablaut) characterizes strong verbs (cf. Modern English sing-sang-sung-song), this is not so for weak verbs. Rather the addition of a dental suffix -d- in the past tense characterizes weak verbs (cf. Modern English arrive-arrived). This dental suffix is appended to the verbal stem, before the addition of personal endings. The dental suffix is found not only in finite verbal forms, but also in the past participle (cf. Modern English 'That problem, addressed by Einstein, was the beginning of modern quantum theory').

WG has four classes of weak verbs that are distinguished by a thematic vowel (thV) which precedes the dental suffix of the preterites, and the presence or absence of a nasal appended to the stem: -i-, -o-, -ái-, -no-. These correspond with different forms of the infinitive.

http://www.utexas.edu/cola/centers/lrc/eieol/gotol-2-X.html#Got02_GP09

Basic Formation of Weak Verbs

The weak verbs stand in contradistinction to strong verbs. Whereas strong verbs employ ablaut to mark the past tense, weak verbs employ a dental suffix -d/þ- (sometimes -t-) to mark the past forms. This formation distinguishes the Germanic family from other branches of Indo-European, as the dental preterite is found nowhere else.

Early investigations into its origin supposed that the dental suffix represents the second element of an original periphrastic formation involving the root *dhē 'put', akin to suggestions that the Latin imperfect represents a periphrastic construction employing the root *bhū 'become'. As with the Latin situation, the hypothesis suffers from a lack of conclusive evidence as to the precise nature of the original construction, e.g. as to what was the form of the first element of the periphrastic construction. Gothic, however, is unique in displaying in the plural a fuller form of the second element than one finds in other Germanic languages. Specifically, while the singular suffixes with their endings monosyllabic: -da, -des, -da, the dual and plural forms are disyllabic: -dēd-u, -dēd-uts; -dēd-um, -dēd-uþ, -dēd-un. These forms are quite suggestive, in that they parallel the attested strong preterite forms of the root *dhē in other Germanic languages: Old High German tāt-um, tāt-ut, tāt-un; Old Saxon dād-un. The disyllabic forms in Wistra-Gutisk are reduced to the point that the -dē- element is completely gone from the paradigm. Past first person singular had the suffix -dēdjáu, it has become -djáu.

Another common thread of investigations into the origin of the dental preterite in Germanic has been its possible relation to the -t- of past participles in Germanic (e.g. WG nas-i-þ-s) and in other branches of Indo-European: Greek do-tó-s 'given'; Latin da-tu-s 'given'. A preterite form with this -t- suffix, such as PIE *kousitōm 'I heard' > PGmc *hauziðōm (EG *hausiðōm), might eventually be conflated in the Proto-Germanic period with PIE *dhōmi 'I put' > PGmc. *ðōm and reanalyzed as a compound form. Then by extension in Gothic the preterite plural forms of PIE *dhē, e.g. PGmc *ðēðum, would have been suffixed to fill out the rest of the paradigm.

Another possibility presents itself, one which does not suffer from the ambiguities of stem formation and analogy which are integral to the preceding theories. Benveniste early studied the use of the dh-determinatve in PIE, especially in the Greek and Indo-Iranian branches. Lehmann (1942, 1943) subsequently illustrated the connotations of this suffix within the Germanic family, and he proposed that this determinative is in fact the origin of the dental preterite. The dh-determinative, sometime in the period leading up to the split of Germanic from PIE, and then subsequently within PGmc, assumed a similar function in three basic situations:

1. with nouns derived from transitive roots: to denote past passive modification. For example, consider PIE *bher- 'cut' > Gothic fotu-baúrd 'foot-board', Old English bord; compare Greek perthō 'destroy', Latin ferīre 'strike', Old High German borōn 'bore'; PIE *mel- 'grind' > Gothic unmidljái (nom. pl. masc.) 'unkind', Old Icelandic mildr 'kind', Old English milde 'kind': the semantic development is evidently 'something that has been ground up' > 'something soft' > 'gentle, mild'. Compare Gothic malan 'grind', Latin molēre 'grind'. Also PIE *wer- 'speak' > Gothic waúrd 'word', Old Icelandic orð 'word', Old English word. Compare Latin verbum 'word', Greek eírō 'speak'. The dh-suffix denotes 'something that has been spoken'.
2. with nouns derived from intransitive roots: to denote modification caused by previous action. For example, consider PIE *gher- 'like', cf. Sanskrit háryati 'likes', Greek khaírō 'rejoice'. The dh-formant survives in Sanskrit grdhyati 'desires', Gothic gredáu (dat. sg. masc.) 'desire', Gothic gredags 'hungry', Old Icelandic gráðr 'hunger'. The semantics pass from 'liking' to the 'result of having liked continuing into the present', and hence 'desiring'. Consider also PIE *men- 'think'. This has a dh-extension PIE *mendh- 'turn one's attention to' > Gothic mundrōn 'turn one's attention to'; compare also Old High German muntar 'alert', Gothic mundrein (dat. sg. fem.) 'desire', resulting from the semantic association 'having turned one's attention towards' > 'alert'.
3. with verbs: to denote modification or change resulting from previous action. For example, consider PIE *(s)keu- 'cover' > Sanskrit skáuti 'covers'. The dh-extension yields Old English hȳdan 'hide, conceal', Greek keúthō 'conceal', as well as the nouns Gothic skauda-raip 'shoe-string', Old Icelandic skjóða 'sack', Middle High German schōte 'covering', Old Persian tigra-xauda 'with pointed cap'. These dh-forms show a semantic development 'to have covered' > 'hidden'. Consider also PIE *ar- 'fit' > Greek ararískō 'arrange', Latin rērī 'think'. The dh-extension gives Gothic undrēdan 'take care of', as well as Old Icelandic ráða, Old English rædan, Old Saxon rādan, all meaning 'give advice'. These show a semantic development 'fitting together, thinking' > 'having thought' > 'giving advice'. Note in addition PIE *wal- 'be strong' > Latin valēre 'be strong'. The dh-extension gives Gothic waldan, Old Icelandic valda, Old English wealdan, Old Saxon waldan, Old High German waltan, Lithuanian veldéti, all 'rule, possess'. The dh-determinative changes 'be strong' to 'have been strong' > 'rule, possess'.

The semantics of the dh-determinative are thus consonant with the eventual dental preterite. Specifically, while the PIE verbal system of aspect was still transparent, the dh-determinative signified a state reached by previous event. As this aspectual system gave way to a tense system, this dh-determinative would have become associated with past action much as the stative developed into the perfect in other branches of IE, such as Greek and Indo-Iranian. Another distinctive feature commending this theory is the simple fact that it takes the elusive nature of the ending of the first periphrastic element at face value: there was none. The original formation is not V1+V2, but rather the typical PIE formation of R+S+E, that is root-suffix-ending.

By the period of the documented Germanic languages the weak verbs are a self-standing pillar of the verbal system. Across the rest of the Germanic languages, these verbs fall into three classes, stemming from the respective suffixation of PIE *-j- > PGmc *-j/ij-, PIE *-ā- > PGmc *-ō-, or PIE *-oi- > PGmc *-ai- to the verbal root. Gothic possesses a fourth class characterized by the suffix PIE *-nō- > PGmc *-nā-. Verbs formed in this manner are found in other Germanic languages, e.g. Old Norse vakna and Old English wæcnian 'awake', but they are not numerous enough to form a class by themselves. In terms of preterite morphology, these verbs conjugate according to the PIE *-ā- > PGmc *-ō- class. Wistra-Gutisk thus possesses the following weak verb classes, usually distinguished in grammars by the form of the infinitive. 

http://www.utexas.edu/cola/centers/lrc/eieol/gotol-8-X.html#Got08_GP37

The First Weak Conjugation

The first weak conjugation is characterized by the suffix PIE *-j- > PGmc *-i/ij- added to the verbal root. These verbs are typically either causative or denominative. Consider the following examples:

In Wistra-Gutisk, as in the other Germanic languages, the reflex of the PGmc *-i/ij- suffix in certain morphological forms depends on the shape of the root to which it was affixed. There are two possibilities: -ji- or -ei- [ī]. First weak conjugation roots thus fall into two groups, based on the reflex of the suffix. The reflex of the suffix only differs between the groups in the 2nd and 3rd person singular present indicative active, in the 2nd person plural present indicative active, and in the 2nd person plural imperative. Elsewhere the form of the suffix does not depend on root shape, but appears as -j- in all other present forms. The suffix appears as -i- in all past forms. For the sake of clarity, paradigms for the two types of Class I weak verbs will be listed separately.

In order to make a clear statement of the distinction between the types of Class I weak verbs, we must refine our terminology of syllable length. Recall that syllables ending in a short vowel are termed short, all others are long. Thus long syllables end in a consonant or contain a long vowel, or both. For the purposes of describing the first weak conjugation, a further distinction must be made. We will restrict the term long syllable and define the term overlong syllable as follows.

• long syllable: a syllable containing a short vowel followed by a consonant, or containing a long vowel with no following consonant;
• overlong syllable: a syllable containing a long vowel followed by a consonant.

Note that traditional terminology in Germanic linguistics often includes the former category under the heading 'short' and terms the latter 'long'. This system has its advantages; the system above is employed here in order to keep terminology as close as possible to that used in grammars of other early Indo-European languages, e.g. Latin, Greek, and Sanskrit.

Class Ia Verbs

Given the above definitions of syllable length, we may succinctly characterize the first type of class i weak verbs:

[Note that this class, according to traditional terminology, contains roots whose stem syllable is 'short' (nasjan) or 'long' and open (stójan).]

The verb nasjan 'to save' serves to illustrate the forms of verbs whose stem contains a short vowel followed by a consonant. The forms are as follows.

For purposes of comparison, the conjugation of stōjan 'to judge' is given below. The forms illustrate the conjugation of Class Ia weak verbs whose stem contains a long vowel without a following consonant.

Class Ia verbs are therefore characterized by the fact that -j- is retained in all present forms, where it is followed by a vowel. Note that the imperative 2 sg. ending is -ei, so the -j- becomes [ī] when final. When -j- comes between two consonants, as in the past forms, it is voiced as -i-. Before a vowel, ō is written au.

*Wistra-Gutisk differs from Gothic with the subjunctive past forms undergoing a reduction of the -dē- element.

Class Ib Verbs

Again according to the above definitions of syllable length, we may succinctly characterize the second type of Class I weak verbs:

[Note that this class, according to traditional terminology, contains roots whose stem syllable is 'long' and closed (sókjan).]

The verb sókjan 'to seek' serves to illustrate the forms of verbs whose stem contains a long vowel followed by a consonant. The forms are as follows.

Class Ib verbs are thus characterized by the fact that -j- is retained before the back vowel -a-, but assimilates with a following -i- to give ī (written ei). As in class Ia the imperative 2 sg. ending is -ei, so the -j- becomes [ī] when final. Likewise, as in class Ia, when -j- comes between two consonants, it is voiced as -i-.

*Wistra-Gutisk differs from Gothic with the subjunctive past forms undergoing a reduction of the -dē- element.

Tense conjugation for Class I Weak Verbs

Wistra-Gutisk gets its various tense conjugation from a blending of Gothic and Icelandic weak paradigms.

In the above chart, the rare usage of munan (to think/consider) is shown. Skal has its own conjugation as a preterite-present verb.

The present perfect tense is constructed using the forms in the chart directly above and the Active Voice preterite forms (indicative or optative). This tense describes an action that happened at an indefinite time in the past or that began in the past and continues in the present. Example: He who has suffered a lot, has learned a lot. Sa-ei habais þuláiðir manags, habais ga-kunnóðir manags.