Abstract
Depending on whether God or a human is the forgiving agent in the Synoptic Gospels (and beyond), the verb valence of ἀφίημι, “forgive,” differs in several ways. The present article argues that the differences are reflections in linguistic conventions of the cognition that only God can remove the substance of sin, while both God and humans can remit the moral debt of sin. Construction grammar, a linguistic theory which assumes that syntax and semantics are inseparable, is used in the analysis. Theological implications are discussed.
1 Introduction
When ἀφίημι is used with the meaning “forgive” in the Synoptic Gospels, the verb takes different accompanying arguments1 depending on whether the one who forgives is God or a human being. I argue that this hitherto overlooked phenomenon can be explained if we assume that two different cognitive frames of forgiveness (imaginations of what forgiveness is) are operative in the Synoptic Gospels, one where sin is a substance that is removed through forgiveness and one where sin is a debt that is remitted through forgiveness. Within each imagination, a distinct usage of ἀφίημι with its own verb valence2 is evoked. Whereas God can be the forgiving agent in both cognitive frames, humans can only be the agent in the latter, which explains why some apparent syntactic possibilities do not turn up when a human being is the forgiving agent. Thus, the linguistic analysis has several theological implications.
2 Statistics on the Valence of ἀφίημι
In bdag, the translation of ἀφίημι as “forgive” is placed under section 2, “to release from legal or moral obligation or consequence.”3 That is, bdag assumes that the event “forgive” in early Christian texts is always modelled on remission of debt—an assumption that this article doubts. From the information provided by bdag, we may infer that ἀφίημι takes three arguments when it means “forgive”:
Agent: the forgiver (nominative case with the verb in the active voice)
Patient:4 the sin/wrong (accusative case with the verb in the active voice)
Beneficiary: the one for whom the <patient> is forgiven (dative case).
A typical example is ἀφῇ ἡμῖν τὰς ἁμαρτίας, “He [agent] forgives the sins [patient] for us [beneficiary]” (1 John 1:9). bdag also notes that sometimes the patient and/or the beneficiary is missing, for example ἀφίενταί σου αἱ ἁμαρτίαι, “Your sins [patient] are forgiven” (Matt 9:2), or, ἀφήσω αὐτῷ; “Shall I forgive to him [beneficiary]?” (Matt 18:21).5 However, neither bdag nor any other linguistic analysis that I have been able to find notices that the verb takes different combinations of arguments depending on whether God or a human being is the agent. Table 1 to 3 show statistics of how the valence of ἀφίημι varies.6 The first number in each cell of the tables is the count in the Synoptic Gospels and the second number the total for the whole New Testament and the Apostolic Fathers. We see several interesting phenomena in the tables, which can be summarized in these five (somewhat overlapping) points:
When God is the agent, there is considerable variation whether the beneficiary, the patient, or both, are mentioned, but when a human is the agent, the by far most frequent case is to mention only the beneficiary.
Only when God is the agent, we have the combination where the patient is mentioned but not the beneficiary.
Almost without exception, the patient is ἁμαρτία only when God is the agent.7
Almost without exception, the patient (usually ἁμαρτία) is only mentioned when God is the agent.8
Only when God is the agent, the verb takes passive form.
3 Can Null Complements Explain the Variation?
Paul Danove, who has researched verb valence in the New Testament extensively, explains the variations in the valence of ἀφίημι, meaning forgive, (as well as many other Greek verbs) with a linguistic phenomenon called “null complements.”28 (“Null complements” are sometimes called “optional arguments” in other research on valence).29 A “null complement” is an argument that is permissible to leave out in a correct sentence if it can be inferred.30 Danove distinguishes two kinds of null complements: A “definite null complement” is an argument that can be left out if it can be derived from its context. An “indefinite null complement,” on the other hand, can be left out even if there is nothing in the context to indicate what it is. In such cases, we infer from general expectations what the implicit argument might be.
According to Danove, when the verb ἀφίημι means “forgive,” the beneficiary is a definite null complement.31 That is, mention of who is forgiven can be left out from the clause if something in the context hints who it is. The patient, on the other hand, is an indefinite null complement. That is, the wrong that is forgiven can be left unmentioned in a correct clause, since the reader can infer what is forgiven from general expectations. As he points out, this is a frequent phenomenon in Greek three-place verbs (verbs with a valence of three arguments).32 Danove’s analysis of three-place verbs has important explanatory value, since it can explain many of the variations recorded in the table above. His analysis will be our starting point. However, the linguistic phenomenon of null complements cannot explain all our observations. For instance, it cannot explain
why the texts avoid explicating the patient when a human is the agent,
why the combination with patient but no beneficiary appears only when God is the agent, and
why the passive voice is used only when God is the agent.
We may suspect that the syntactic differences indicate that divine forgiveness was somehow perceived as qualitatively different from human forgiveness. The problem, I argue, is that Danove (just like bdag) assumes that ἀφίημι is always a three-place verb when it means “forgive.”
4 Two Cognitions of Forgiveness: Removing a Substance and Releasing a Debt
Gary Anderson has recently analysed how Jews and Christians have understood sin as analogous to substance and debt in his monograph Sin: A History.33 With the aid of conceptual metaphor theory,34 he argues that in pre-exilic texts sin is thought of as a substance, either a burden or a stain. His most thorough analysis concerns the Hebrew word, נשׂא, “forgive,” which literally means “lift, carry.” The expression נשׂא עון can mean both “forgive [i.e. carry away] sin” and “bear (punishment for) sin.” Such linguistic peculiarities are only possible if sin and forgiveness are cognitively modelled on the concrete human experience of burdens.35
Anderson then argues that during the exile there was a major shift in how Jewish texts talk about sin. When Jews came into contact with Aramaic, they adopted Aramaic idioms for sin, which were modelled on monetary debt. For instance, Aramaic Targums consistently translate נשׂא עון to שׁבק חובה, “remit debt,” Anderson points out.36 This, in turn, not only changed how they imagined sin but also the remedies for sin. The imagination of forgiveness changed from removing a substance to remitting a debt. With the cognitive frame of debt, new ways of thinking about sin developed, such as the idea that one could pay the debt of sin through suffering and good deeds.37
Anderson’s analysis is compelling, and my argument builds on his results. There is, however, one important modification I would like to make. Anderson claims that the metaphor of substance “was replaced” by the metaphor of debt.38 I would say that a more accurate description of the development is that the metaphor of debt became prevalent, but without eradicating the metaphor of substance. In post-exilic Judaism, including early Christianity, both the cognitive frame of substance and the cognitive frame of debt were used to understand sin and forgiveness. For instance, the post-exilic language of sin as something that renders the sinner morally unclean is unintelligible unless the substance-imagination was alive and well in Jewish cognition (e.g. Sir 23:10; 38:10; 1 Macc 1:48; Philo, Cher. 28:91-95; Jub. 9:15; 1qs 3:13-14).39 The idea of sin as substance, which needs to be lifted, cleansed, or covered, was also part of early Christian thinking and language on sin and forgiveness (e.g. John 1:29; 2 Tim 3:6; Heb 1:3; 9:28; 10:2, 4, 11; 12:1, 4; Jas 3:6; 1 Pet 4:8; 2 Pet 1:9; 1 John 1:7, 9; Rev 7:14; 22:14). In the Synoptic Gospels, sin is perceived of as substance when Jesus claims that moral evil defiles the sinner (Mark 7:20-23; Matt 15:18-19).
5 Why Did the Translators of the Septuagint Choose ἀφίημι?
In extra-Biblical Greek, ἀφίημι is not used to talk about forgiveness. lsj does not even mention “forgive” as a possible translation for ἀφίημι.40 Bauer-Aland manages to find a limited number of non-Jewish examples, though.41 The word that comes closest to “forgive” in extra-Biblical Greek is συγγιγνώσκω.42 Therefore, the Christian usage of ἀφίημι to say “forgive” is best explained by the fact that the Septuagint frequently choose to translate סלח and נשׂא, the two most common Hebrew words for “forgive,” with ἀφίημι.43
Why did the translators of the Septuagint choose ἀφίημι? We will never know. However, the general tendency of “semantic borrowing” in the Septuagint is pedagogically described by Karen Jobes and Moisés Silva:
The process [of semantic borrowing] is fairly clear: speakers first notice some semantic correspondence between a word in their language and a similar word in the foreign language, then proceed to bring the usage . . . of the two closer together. . . . semantic borrowing involves extending the area covered by one word so that the overlap becomes greater or even complete.44
Still, why not choose συγγιγνώσκω, like Josephus (e.g. Ant. 2.145, 154; 3.23; 6.151, 219, 303)? Why not συγγνώμη or ἀμνηστία together with a fitting verb, like Philo (e.g. Spec. 1.42, 229, 235-236, 242; 2.23; 3.121)?45 At first glance, there seems to be little reason to choose ἀφίημι.
As we can only speculate about why the translators chose ἀφίημι, I will allow myself to do so. Suppose the translators of the Septuagint perceived forgiveness of sin both as removing a substance and as relieving a debt. If so, they would have to find a Greek word that could accommodate both these cognitive frames—and ἀφίημι, in its astonishing polysemy, is able to do just that with a little bit of semantic borrowing. ἀφίημι can mean “send away (an object),” which fits the cognition of removing the substance of sin. It can also mean “remit (a debt),” which fits the cognition of forgiving the debt of sin.
6 A Construction Grammar Analysis of the Conventional Greek Use of ἀφίημι
Our next step is to use the conventions developed in construction grammar to describe the syntactic and semantic qualities of ἀφίημι in a way that helps us understand how the verb can function to denote “send/let an object on a trajectory” and “remit a debt for someone.”46 The fundamental idea of construction grammar is that syntax and semantics are not separable in real language. Rather, language consists of “constructions” which have both semantic and syntactic properties.47
One of the basic assumptions of construction grammar is “The Principle of No Synonymy,” which means that “If two constructions are syntactically distinct, they must be semantically or pragmatically distinct.”48 For instance, the verb “drink” is part of several syntactically, semantically, and pragmatically, distinct constructions, such as: “I drink water,” where the beverage needs to be mentioned; “I do not drink,” where some kind of alcoholic beverage is implied but should be left unmentioned; and “I drink you under the table,” where a drinking-fellow and the expression “under the table” is expected to accompany the verb. With this approach to language, a word like ἀφίημι, in all its polysemy and polyvalence, cannot be a construction by itself, but is part of many different constructions, which evoke many different cognitive frames. In our analysis of ἀφίημι, the important aspects of each construction to describe are a) semantic frame elements (different parts of the cognition evoked in our minds by the construction), and b) valence.49
Construction: “A sender sends/lets an object on a trajectory”
Citation: Novum Testamentum 57, 2 (2015) ; 10.1163/15685365-12341489
According to lsj, one basic meaning of ἀφίημι with two arguments (agent and patient) is “send forth” and “send away.”50 lsj also suggests several other ways to translate the word in different contexts, for instance “put forth,” “discharge,” “let loose,” and “let go.” All these usages of the word evoke the basic event “a sender sends/lets an object on a trajectory” and apply it by metaphorical extension to different kinds of events.51 The construction is described in figure 1.
A few explanatory notes: A “head lexeme” is the lexical form of the verb functioning as predicate and thus governing the construction. In the section called the “semantic frame elements,” I have described the most important (but not all) frame elements that are evoked in our cognition when we think about the event “send/let an object on a trajectory.” All of these elements may be part of a correct and meaningful sentence, but only the arguments enumerated under the section “valence” are necessary for making the sentence a meaningful one. The numbers connect the different semantic roles to the different frame element. In this case, the semantic role agent is the sender, and the semantic role patient is the sent object.
In the lexical entry of lsj, the usage of ἀφίημι in the sense “remit” is hidden away under section A.II.2.c. There it is rightly noted that the valence is different for this usage: “c. dat. pers. et acc. rei.” That is, the verb now has a third argument, a beneficiary. Moreover, the cognitive frame evoked by this usage is quite distinct from other usages in section A.II of the entry. Had the entry been arranged according to the principles of construction grammar, the specialized meaning together with the differing valence would have merited this usage a new section in the entry.52 bdag, on the other hand, has a new section for this usage.53 The construction is described in figure 2.
Construction: “A benefactor remits a bond/debt/obligation for a person”
Citation: Novum Testamentum 57, 2 (2015) ; 10.1163/15685365-12341489
Thus we have two different usages of ἀφίημι, which with a little bit of semantic borrowing (see above) can accommodate the two different cognitions of forgiveness of sin. The construction “A sender sends/lets an object on a trajectory” can be used to describe removal of the substance sin, and the construction “A benefactor remits a bond/debt/obligation for a person” can be used to describe the remission of the debt of sin. The remainder of the present article explores how this distinction applies to a number of key passages in the Synoptic Gospels in order to describe and interpret, as adequately as possible, the constructions in which ἀφίημι appears with the meaning “forgive.”
7 The Story of the Paralytic Whose Sins are Forgiven
In the story of healed paralytic whose sins are forgiven in Mark 2:1-12, ἀφίημι is consistently used without the semantic role beneficiary. In the two passive formulations, ἀφίενταί σου αἱ ἁμαρτίαι, (2:5, 9), the agent is missing, and the only explicitly mentioned argument is the patient, σου αἱ ἁμαρτίαι.54 In the two active formulations, τίς δύναται ἀφιέναι ἁμαρτίας εἰ μὴ εἷς ὁ θεός; (2:7) and ἐξουσίαν ἔχει ὁ υἱὸς τοῦ ἀνθρώπου ἀφιέναι ἁμαρτίας (2:10), the semantic roles agent and patient are spelled out.
No Beneficiary Argument and No Omission of the Patient Argument
If ἀφίημι is used to talk about removing sin, we should expect a patient but no beneficiary. Thus, it is reasonable to assume that the story works with a concept of forgiveness as removal of substance rather than remission of debt, since the pericope consistently uses ἀφίημι without mentioning a beneficiary argument. This impression is further strengthened if we look at the context and content of the narrative.
The narrative is part of a range of pericopae in the beginning of the Gospel of Mark that together serve to show Jesus’ ἐξουσία, “authority/power” (1:21-2:12, esp. 1:22, 27; 2:10). The plot of the story is whether Jesus has the capacity (δύναται, 2:7) to do what only God can do. As the statistics above show, God is normally the agent in the combination where the patient is mentioned but not a beneficiary, which evokes the substance-frame of forgiveness. The question “Who can forgive sin but one, God?” is intelligible only if it evokes the cognitive frame of forgiveness that is uniquely linked with God in early Christian literature—removal of the substance of sin.
The story also plays on the widely held belief that there was a connection between sin and disease.55 That Jesus can remove the bodily sickness proves that Jesus has also removed its cause, sin. In Jewish thinking, sin sometimes causes God to punish with sickness (e.g. Exod 20:5; Lev 26:14-33; Deut 28:15-16; 2 Chr 12:12-19). Sometimes sin itself becomes a destructive power or opens way for a demonic power that causes sickness and moral weakness (e.g. Ps 38:4; 40:12-13; Prov 5:22; Wis 11:15-16; Rom 5:12-8:39; 11QPsa 19:13-16). Both these cognitions are compatible with the idea of sin as dangerous substance that needs to be removed. The play on a connection between sin and sickness in the story therefore strengthens the case that the valence of ἀφίημι in this passage is used to evoke the removal-frame.
The impression from Mark is intact in Matthew’s version of the story (9:1-8), which largely preserves Mark’s version.56 Just like Mark, Matthew places the story in a literary context where Jesus’ ἐξουσία is central. At the end of the Sermon of the Mount, people are amazed by his ἐξουσία (7:28-29). The sermon is followed by a number of pericopae where Jesus exercises this power (8:1-9:35). Thereafter the disciples are given ἐξουσία (10:1).
One fascinating modification in Matthew is worth attention, though. The story ends with praise of “God who has given such authority to humans” (τὸν θεὸν τὸν δόντα ἐξουσίαν τοιαύτην τοῖς ἀνθρώποις, 9:8). This verse probably hints the same idea as Matthew 18:15-20. There, the Matthean community is entrusted with the authority to make authoritative halakhic decisions and mediate divine forgiveness (esp. 18:18).57 The Matthean text consistently avoids describing interpersonal forgiveness as removal of a substance (6:12, 14-15; 8:20-35), so when this text talks about human authority to forgive sin in the substance-frame, it probably extends the authority given to Jesus in Markan version to the community of his followers.58
In the story of the forgiven paralytic, the patient argument is consistently spelled out and never omitted. According to Danove’s analysis (presented above), it is allowed to leave out the patient argument for predicates with three arguments, since it can be inferred from general expectations (indefinite null complement). This is only true when forgiveness is modelled on remission of debt, which is expressed with a linguistic construction with three arguments. When forgiveness removes the substance of sin, there is not one example in early Christian literature where the patient argument is omitted. Moreover, the patient is without exception ἁμαρτία (Mark 2:5; 2:7; 2:9; Matt 9:2; 9:5; Luke 5:21; 7:47a; 7:48; John 20:23; Rom 4:7; 1 Clem. 50:6; Herm. Sim. 7.1.4; Did. 11:7). These observations strongly suggest that we are dealing with a linguistic construction which is distinct from that of debt frame.
In sum, what we have described is a construction with two arguments, agent and patient. Within usage-based construction grammar, the goal is to describe as accurately as possible how a construction is used in actual language.59 Therefore, we suggest that in the prototypical usage of this construction, there are not only syntactic but also semantic expectations: a) the agent should be God, and b) the patient should be ἁμαρτία.
Redundancy in Luke
Luke modifies Mark’s two usages of ἀφίημι in the passive form. He conflates the two possibilities ἀφέωνταί αἱ ἁμαρτίαι σου and ἀφέωνταί σοι αἱ ἁμαρτίαι into ἀφέωνταί σοι αἱ ἁμαρτίαι σου (5:20, 23).60 That is, he adds a redundant beneficiary argument, σοι, to Mark’s text without removing the genitive attribute, σου, which also functions to tell us who is forgiven. This double indication of who is forgiven recurs a few more times in early Christian texts, so Luke is not alone to do this (Mark 11:25; Matt 6:12 // Luke 11:4 // Did. 8:2; Acts 8:22; 1 Clem. 60:1; a few text variants of Matt 9:2).61 It is also difficult to know why he changes Mark here, considering that he uses the construction without beneficiary in 7:47a (ἀφέωνται αἱ ἁμαρτίαι αὐτῆς) and 7:48 (ἀφέωνταί σου αἱ ἁμαρτίαι), which is unique material for Luke.
The phenomenon could possibly be explained by the linguistic phenomenon called “contamination” or “blending,” where two syntactically and semantically similar constructions are fused into a new construction.62 If so, the existence of these examples of contamination is a further indication that Danove’s explanation is not the whole truth. Rather, an unnecessary double indication of whose sins are forgiven could be the result of a blend of two separate but related ways to talk about forgiveness—one where the forgiven person is indicated with a genitive attribute to the patient argument, and one where the forgiven person has the semantic role of beneficiary.
Another possible explanation begins with the observation that redundant personal pronouns were quite common in Koine Greek. One of the functions of such redundancy was to create emphasis, but Chrys Caragounis has recently demonstrated that redundant pronouns, especially pronouns in the genitive case, appear frequently in the New Testament with no apparent function.63 If so, we may suspect that Luke was influenced by the redundant use of a genitive pronoun in the Lord’s Prayer (ἄφες ἡμῖν τὰς ἁμαρτίας ἡμῶν, Luke 11:4 // Matt 6:12 // Did. 8:2), a prayer that Luke quite possibly knew by heart. The construction in the Lord’s Prayer is clearly in the debt frame (see analysis below), which means that Luke might have imported a syntax which is not fully sensitive to the narrative’s understanding of what sin is.
The Use of the Passive Voice
In early Christian texts, the passive form of ἀφίημι is normally reserved for clauses where God is the agent, as the statistics above show. In all these cases, the agent argument is not mentioned, but the reader is supposed to be able to infer from context and linguistic conventions that God is the agent. This is the well-known linguistic phenomenon of passivum divinum.64 Many verbs that are sometimes used to express passivum divinum are also on other occasions used in the passive with human agents. ἀφίημι, however, is never used in the passive form in discourse about interpersonal forgiveness. The passive form is reserved for divine agency in early Christian language conventions. We conclude that in early Christian discourse, the prototypical use of ἀφίημι in the passive voice is a) without explicit agent, but b) with God as the implicit agent.
In the story about the forgiven paralytic, the interpretation of the passive forms of ἀφίημι as passivum divinum becomes a bit more problematic. Tobias Hägerland has argued that the passive formulation, ἀφίενταί σου αἱ ἁμαρτίαι (2:5, 9), goes back to the historical Jesus and that the historical Jesus saw himself as a mediator of God’s forgiveness.65 The passive utterance would then be an unobjectionable passivum divinum. However, in the story told by Mark, the passive formulation causes confusion and accusation that Jesus blasphemes. Beniamin Pascut has recently argued that passivum divinum is a too simple interpretation of the passives in 2:5, 9.66 The plot of the pericope in Mark is not intelligible if Jesus is just stating what God has done, Pascut argues. Rather, the statements function performatively to effect forgiveness. If the passive is a performative utterance, we can understand both why the scribes become upset (2:6-7) and why the story concludes that the Son of Man does indeed have authority to forgive sins (2:10).
Is this authority independent of God’s? The verb ἔχω in the phrase ἐξουσίαν ἔχει ὁ υἱὸς τοῦ ἀνθρώπου opens up for the possibility that Jesus has authority independent of God’s authority. Given the theology of Mark where Jesus is portrayed as God’s agent of the kingdom of God (e.g. 1:11, 15; 9:7; 10:18, 45), Morna Hooker’s interpretation is more reasonable. She suggests that Mark intends to say that authority of the Son of Man is dependent on God’s, just like the authority of the “one like a human being” in Dan 7:13-14 is given by God.67 Pascut allows that—theologically—God may still be considered the ultimate agent behind the miraculous actions performed by Jesus in the Gospel of Mark, including the forgiveness of sins. We may therefore still consider the passive formulations in 2:5, 9 to be an indirect divine passive, since God is the ultimate power behind the efficacy of the speech-act.68 Mark skilfully plays on the linguistic convention of passivum divinum to demonstrate Jesus’ God-endowed authority.
8 The Lord’s Prayer
All preserved versions of the Lord’s Prayer exhibit a peculiar syntactic asymmetry between how divine and human forgiveness is expressed. In the first clause, where God is the agent, both patient and beneficiary are mentioned. In the second clause, where the agent is human, the patient is omitted and only the beneficiary is referenced.
ἄφες ἡμῖν τὰ ὀφειλήματα ἡμῶν, ὡς καὶ ἡμεῖς ἀφήκαμεν τοῖς ὀφειλέταις ἡμῶν
(Matt 6:12)
ἄφες ἡμῖν τὴν ὀφειλὴν ἡμῶν, ὡς καὶ ἡμεῖς ἀφίεμεν τοῖς ὀφειλέταις ἡμῶν
(Did. 8:2)
ἄφες ἡμῖν τὰς ἁμαρτίας ἡμῶν, καὶ γὰρ αὐτοὶ ἀφίομεν παντὶ ὀφείλοντι ἡμῖν
(Luke 11:4)
Limits to the Analogy between Divine and Human Forgiveness
Both human and divine forgiveness is here clearly modelled on the debt-metaphor. In Matthew and the Didache, the metaphor for sin is ὀφείλημα/ὀφειλή, ”debt,” and in all three versions the beneficiary of human forgiveness is an ὀφειλέτης, “debtor.”69 In Matthew and the Didache ὡς indicates that divine and human forgiveness are analogous. The καὶ γάρ in Luke signals that interpersonal forgiveness is a reason for God’s forgiveness. A survey of all instances in the New Testament and the Apostolic Fathers that express some kind of connection between God’s forgiveness and interpersonal forgiveness show that all but one (John 20:23, discussed below) are modelled on the debt-metaphor, which is indicated by an explicit mention of the beneficiary (Mark 11:25; Matt 6:12, 14-15; 18:21-35; Luke 11:4; Did. 8.2; 1 Clem. 13.2; Pol. Phil. 2:3; 6:2). That is, wherever God’s forgiveness motivates interpersonal forgiveness, the process of forgiveness is described as remission of debt.
Nonetheless, all versions of the prayer for forgiveness in the Lord’s Prayer mention the patient when God is the agent but not when the agent is human. (This asymmetry can also be found in Mark 11:25.) Together with the general pattern that the patient is almost never mentioned when humans are forgiving agents (see the statistics above), these asymmetries indicate that in early Christian linguistic culture the nature of what was forgiven (the patient) was usually not specified in discourse about interpersonal forgiveness. We can only speculate as to why this linguistic convention developed, but perhaps the intuition that only God can forgive sin inspired speakers to leave out the patient argument.70 Nevertheless, we may claim that in the prototypical use of ἀφίημι in discourse about interpersonal forgiveness, the patient argument should be left out.
One passage in Matthew partly deviates from the overall pattern just described. When Matthew wishes to reinforce the importance of interpersonal forgiveness after the Lord’s Prayer, he formulates the following.
a. Ἐὰν γὰρ ἀφῆτε τοῖς ἀνθρώποις τὰ παραπτώματα αὐτῶν,
b. ἀφήσει καὶ ὑμῖν ὁ πατὴρ ὑμῶν ὁ οὐράνιος•
b.’ ἐὰν δὲ μὴ ἀφῆτε τοῖς ἀνθρώποις,
a.’ οὐδὲ ὁ πατὴρ ὑμῶν ἀφήσει τὰ παραπτώματα ὑμῶν.
(6:14-15)
Here the patient, τὰ παραπτώματα αὐτῶν, is mentioned in a clause about interpersonal forgiveness. Content-wise the first half is an antithetical parallelism to the other half. Syntactically, however, the sentence forms a chiasm, where a and a’ mentions τὰ παραπτώματα, but not b and b’. We may therefore assume that the striving for a symmetric sentence structure got the upper hand over linguistic conventions in this particular case. Nevertheless, interpersonal forgiveness is still modelled on remission of debt, since the beneficiary is mentioned. Moreover, Matthew chooses the word παράπτωμα—a word that is used only here in Matthew—probably in order to avoid the word ἁμαρτία. (All passages in Matthew that expresses some kind of connection between divine and human forgiveness use the debt-frame and avoid the term ἁμαρτία, 6:12, 14-15; 18:23-35.)
A Brief Note on John 20:23
Although this paper focuses on the Synoptic Gospels, a brief mention of John 20:23 is proper here. Having breathed the Spirit over the disciples in the preceding verse, Jesus promises that ἄν τινων ἀφῆτε τὰς ἁμαρτίας ἀφέωνται αὐτοῖς, “If you forgive the sins of any, they are forgiven to them.” In the conditional clause, which is about human forgiveness, the evangelist mentions the patient but not the beneficiary. Moreover, the patient is ἁμαρτία. This construction is otherwise only used when God is the agent. As argued above, this syntax evokes the cognitive frame of forgiveness as removal of substance, which normally only God is able to effectuate. Many scholars have tried to avoid the theologically difficult conclusion that the Johannine community actually considered itself authorized to remove sin. In this scholarly discussion, a syntactical problem has been debated: Does the perfect tense of ἀφέωνται in the main clause mean that God’s forgiveness comes first so that the community only consents, or does the just mentioned reasoning not apply in relation to conditional clauses?71 The results of this paper add another syntactic argument to the discussion: The valence of the clause about human forgiveness fits perfectly into the substance-frame, which is usually only used when God is the agent. The linguistic construction is therefore unique in early Christian literature. This indicates that the statement is precisely as theologically provocative as it seems to be.72
9 The Saying on Interpersonal Forgiveness in Mark 11:25
Mark 11:25 has a rare usage of ἀφίημι, with neither patient nor beneficiary. The same usage can be found once in 1 Clement and twice in Polycarp’s Letter to the Philippians:
ἀφίετε εἴ τι ἔχετε κατά τινος, ἵνα καὶ ὁ πατὴρ ὑμῶν ὁ ἐν τοῖς οὐρανοῖς ἀφῇ ὑμῖν τὰ παραπτώματα ὑμῶν
(Mark 11:25)
ἀφίετε ἵνα ἀφεθῇ ὑμῖν
(1 Clem. 13:2)
ἀφίετε καὶ ἀφεθήσεται ὑμῖν
(Pol. Phil. 2:3)
εἰ οὖν δεόμεθα τοῦ κυρίου ἵνα ἡμῖν ἀφῇ, ὀφείλομεν καὶ ἡμεῖς ἀφιέναι
(Pol. Phil. 6:2)
In all these passages, the usage of the verb with neither patient nor beneficiary is possible because its clause is paralleled by a clause that includes at least the beneficiary argument. That is, the parallel clause evokes the frame of sin as remission of debt. Moreover, the wider contexts of all these statements allow the reader to infer who the beneficiary is. This is a good example of where Danove’s explanation (presented above) is very helpful: in the debt-frame, the patient may be omitted and inferred from general expectations (indefinite null complement) and the beneficiary may be omitted if given by the context (definite null complement).73
10 The Story of the Sinful Woman in Luke 7:36-50
The story of the sinful woman who throws herself at Jesus’ feet and receives forgiveness appears only in Luke (7:36-50). In vv. 47-48, ἀφίημι is used in a way that goes back and forth between the substance-frame and the debt-frame. Could this passage possibly undermine the argument of this paper? On the contrary, at closer inspection it is the other way around.
The ending of the story is confusing. In the dialogue between Jesus and Simon, the point is that forgiveness causes a response of love (7:40-46). Then the causal connection is reversed, so that love merits forgiveness (7:47a); then back again to the idea that forgiveness causes love (7:47b); then a final reversal of the causality when Jesus seems to affirm that the woman’s acts of love together with her faith has merited forgiveness (7:48-50). This confusion about what causes what has led to a mountain of interpretations, and quite a few scholars conclude that the Lukan text intertwines different traditions without fully smoothing out the tensions.74
I have no intention to solve the problem here, but if we concentrate on how ἀφίημι is used, we find an interesting contribution to this scholarly debate, which also at the same time might explain why Luke mixes two different ways of talking about forgiveness. In the dialogue, where forgiveness causes love, the text evokes the debt-frame (ἀμφοτέροις ἐχαρίσατο, 7:42; ᾧ τὸ πλεῖον ἐχαρίσατο, 7:43). The verb χαρίζομαι is used here rather than ἀφίημι, probably because the parable that Jesus tells is about real monetary debt.75 In 7:47a, where love causes forgiveness, the text uses ἀφίημι with a syntax that evokes the substance frame (ἀφέωνται αἱ ἁμαρτίαι αὐτῆς αἱ πολλαι). Then, in 7:47b, which returns to the idea that forgiveness causes love, he also returns to the syntax typical of the debt-frame (ᾧ δὲ ὀλίγον ἀφίεται). Finally, in 7:48-50, where the text seems to return to the idea that acts of love causes forgiveness, the syntax of forgiveness changes again to that typical of the substance-frame (ἀφέωνταί σου αἱ ἁμαρτίαι, 7:48; ἁμαρτίας ἀφίησιν, 7:49). That is, the idea that love causes forgiveness is consistently paired with the substance-frame of forgiveness, but the idea that forgiveness causes love is consistently paired with the debt-frame of forgiveness. This gives reason to suggest that Luke does indeed intertwine material from at least two sources. When Luke attempts dialectic integration of two understandings of the relation between love and forgiveness, he also intertwines two different linguistic constructions for forgiveness.
11 Conclusion
When early Christians used ἀφίημι in discourse about forgiveness, they used not one but two different linguistic constructions, which evoked distinct semantic and syntactic expectations. In one construction, removal of substance is the conceptual metaphor upon which forgiveness is modelled. To forgive is to remove of the substance of sin. Only God can be the agent. The expected valence is two arguments, agent and patient. A full description of this construction is shown in figure 3.
In the other construction, remission of debt is the conceptual metaphor upon which forgiveness is modelled. To forgive is to remit the debt of sin. Both God and humans can be the forgiving agent, but when humans forgive, the patient should be left unmentioned. A full description of this construction is shown in figure 4.
The peculiar syntactic properties of ἀφίημι thus reflect the complexity of the early Christian imagination of forgiveness and sin. Only God had the power remove the substance of sin, but humans both could and should forgive the moral debt of those who wronged them, just like God. That is, God’s benevolence rather than God’s power patterned interpersonal forgiveness. More controversial is whether the linguistic observations of this paper can be used to argue that God’s power to forgive the substance of sin was considered to be extended to the church in the in Johannine and Matthean tradition (John 20:23; Matt 9:8), but the results of our attention to verb valence support this interpretation.
Construction: “A divine agent sends away/forgives the substance of sin”
Citation: Novum Testamentum 57, 2 (2015) ; 10.1163/15685365-12341489
Construction: “A benefactor remits/forgives the debt of sin to a sinner”
Citation: Novum Testamentum 57, 2 (2015) ; 10.1163/15685365-12341489
1 For a linguistic statement to be complete, the predicate (a verb) needs to be accompanied by the correct number of arguments (also called complements). Each argument has semantic role (also called thematic role or case role) in relation to the predicate. For instance “build” needs two arguments, 1) an agent (who builds) and, 2) a patient (that is built); e.g. “Lisa [agent] builds [predicate] a house [patient].” See e.g. B. Aarts, English Syntax and Argumentation (2nd ed.; Basingstoke: Palgrave, 2001) 91-97. In general, I have used the linguistic terms preferred in Mirjam Fried and Jan-Ola Östman’s introduction to construction grammar, M. Fried and J.-O. Östman, “Construction Grammar: A Thumbnail Sketch,” in Construction Grammar in a Cross-Language Perspective (ed. M. Fried & J.-O. Östman; Amsterdam: John Benjamins, 2004) 11-86.
4 We could also say that this argument has the semantic role theme, but I have chosen patient for the sake of simplicity. The semantic role patient is defined as “the ‘undergoer’ of an action” while theme is defined as the “the entity that is moved by the action,” according to for instance the textbook of Aarts, English Syntax, 95. However, scholars disagree on how and if these two semantic roles should be distinguished, see e.g. B.J. Blake, Case (2nd ed.; Cambridge: Cambridge up, 2001) 70-71; D. Dowty, “Thematic Proto-Roles and Argument Selection,” Language 67 (1991) 547-619.
In New Testament scholarship working with semantic roles, S.S.M. Wong, A Classification of Semantic Case-Relations in the Pauline Epistles (New York: P. Lang, 1997) 244, suggests that this argument of ἀφίηµι should be considered a patient. Paul L. Danove also suggests “patient” in Linguistics and Exegesis in the Gospel of Mark: Applications of a Case Frame Analysis (Sheffield: Sheffield Academic 2001) 160. However, in his later work, A Grammatical and Exegetical Study of New Testament Verbs of Transference: A Case Frame Guide to Interpretation and Translation (Library of New Testament Studies. London: Continuum, 2009) 131-132, he subsumes the role patient under the role theme.
5 The translations of the beneficiary argument in this paper are sometimes somewhat clumsy on purpose, since more idiomatic translations to English like “He forgives us our sins” gives the impression that the forgiven person has the semantic role of patient.
6 The following occurrences are left out of the statistics, since it is debatable whether the agent should be considered human or divine: Matt 18:27; 18:32, a parable where the metaphorical agent is a human king in the source/vehicle but God in the target/tenor; Mark 2:10; Matt 9:6; Luke 5:24; 7:49, where Jesus, the intermediary of God, is the agent. The latter texts will be analysed below.
9 Mark 3:28; 11:25; Matt 6:12; 12:31; Luke 5:20; 5:23; 11:4; Acts 8:22; 1 John 1:9; 2:12; 1 Clem. 50:5; 53:4; 60:1; Herm. Vis. 2.2.4; Did. 8:2.
11 Mark 2:5; 2:7; 2:9; Matt 6:15; 9:2; 9:5; Luke 5:21; 7:47; 7:48; Rom 4:7; 1 Clem. 50:6; Herm. Sim. 7.1.4; Did. 11:7 (2x).
13 Mark 4:12; Matt 6:14; 12:32; Luke 7:47; 12:10; [23:34]; John 20.23; Jas 5:15; 1 Clem. 13:2; 51:1; Ign. Phld. 8:1; Pol. Phil. 2:3; 6:2.
16 Mark 2:5; 2:7; 2:9; 3:28; Matt 9:2; 9:5; 12:31; Luke 5:20; 5:21; 5:23; 7:47; 7:48; 11:4; Jas 5:15; 1 John 1:9; 2:12; 1 Clem. 50:5; 53:4; Herm. Vis. 2.2.4; Herm. Sim. 7.1.4; Did. 11:7 (2x).
22 Rom 4:7; 1 Clem. 50:6; 60:1. (In 1 Clem. 60:1, the patient is actually τὰς ἀνοµίας ἡµῶν καὶ τὰς ἀδικίας καὶ τὰ παραπτώµατα καὶ πληµµελείας.)
23 Mark 4:12; Matt 6:14; 12:32; Luke 7:47b(?); 12:10; [23:34]; John 20:23; Jas 5:15; 1 Clem. 13:2; 51:1; Ign. Phld. 8:1; Pol. Phil. 2:3; 6:2. (Luke 7:47b, ᾧ δὲ ὀλίγον ἀφίεται, is a difficult case. Since the adjective ὀλίγον is neuter and has no definite article, ὀλίγον could be an adverbial expression, an attribute to an implicit but unmentioned patient in the neuter such as παράπτωµα, or even the patient. However, since it is neuter it is probably adverbial, cf. Mark 1:19; 6:31; 1 Pet 1:6; 5:10; Rev 17:10.)
24 Mark 11:25; Matt 6:12; 6:15; 18:21; 18:35; Luke 11:4; 17:3; 17:4; 1 Clem. 13:2; Pol. Phil. 2:3; 6:2; Did. 8:2.
25 Mark 2:7; 11:25; Matt 6:12, 6:14; 6:15; Luke 5:21; 11:4; [23:34]; 1 John 1:9; 1 Clem. 53:4; 60:1; Ign. Phld. 8:1; Pol. Phil. 2:3; 6:2; Did. 8:2.
26 Mark 11:25; Matt 6:12; 6:14; 6:15; 18:21; 18:35; Luke 11:4; 17:3; 17:4; John 20:23; 1 Clem. 13:2; Pol. Phil. 2:3; 6:2; Did. 8:2.
27 Mark 2:5; 2:9; 3:28; 4:12; Matt 9:2; 9:5; 12:31; 12:32; Luke 5:20; 5:23; 7:47 (2x); 7:48; 12:10; John 20:23; Acts 8:22; Rom 4:7; Jas 5:15; 1 John 2:12; 1 Clem. 13:2; 50:5; 50:6; 51:1; Herm. Vis. 2.2.4; Herm. Sim. 7.1.4; Did. 11:7 (2x).
29 E.g. G. Helbig and W. Schenkel, Wörterbuch zur Valenz und Distribution deutscher Verben (8th ed.; Tübingen: Niemeyer, 1991) 31-39, use the term “fakultative Valenz.”
30 Danove uses the terminology of C.J. Fillmore “Pragmatically Controlled Zero Anaphora,” in Proceedings of the Twelfth Annual Meeting of the Berkeley Linguistics Society (ed. N. Nikiforidou et al.; Berkeley, Calif.: Berkeley Linguistics Society, 1986) 95-107.
33 G. Anderson, Sin: A History (New Haven: Yale up, 2009). Anderson is not the first to observe that sin understood as substance or debt in Jewish and Christian thinking. Rather, much of his basic insights can be found in standard exegetical dictionaries, e.g. M.G. Vanzant, “Forgiveness,” nidb 2 (2007) 480-485; J.S. Kselman, “Forgiveness,” abd 2 (1992) 831-833. However, since Anderson uses cognitive semantics in his analysis, he is able to show the implications of these observations better than previous studies.
34 G. Lakoff and M. Johnson, Metaphors We Live By (Chicago: University of Chicago Press, 1980); G. Lakoff and M. Johnson, Philosophy in the Flesh: The Embodied Mind and Its Challenge to Western Thought (New York: Basic Books, 1999).
39 Cf. T. Kazen, Issues of Impurity in Early Judaism (Winona Lake: Eisenbrauns, 2010) esp. 13-40; J. Klawans, Impurity and Sin in Ancient Judaism (New York: Oxford up, 2000).
41 W. Bauer, Griechisch-deutsches Wörterbuch zu den Schriften des Neuen Testaments und der frühchristlichen Literatur (ed. K. Aland and B. Aland; 6th ed.; Berlin: de Gruyter, 1988) s.v. ἀφίηµι.
42 However, David Konstan has argued convincingly that συγγιγνώσκω matches neither ancient Jewish concepts of forgiveness, nor our modern ideas of forgiveness (Before Forgiveness: The Origins of a Moral Idea [Cambridge: Cambridge University Press, 2010] 22-59).
43 Note, however, that the Pauline tradition prefers χαρίζοµαι to ἀφίηµι (2 Cor 2:7, 10; 12:13; Eph 4:32; Col 2:13; 3:13).
44 K.H. Jobes and M. Silva, Invitation to the Septuagint (Grand Rapids, Mich.: Baker Academic, 2000) 108-109; cf. M. Silva, Biblical Words and Their Meaning: An Introduction to Lexical Semantics (rev. and expanded ed.; Grand Rapids, Mich.: Zondervan, 1994) 53-100.
45 Philo uses ἀφίηµι in the sense “forgive” in a few direct quotes from the Septuagint, e.g. Her. 1.20; Det. 1.141. Josephus occasionally uses ἀφίηµι in the sense “forgive” or “pardon,” e.g. A.J. 2.146; 6.92. However, Josephus uses the construction “c. acc. pers. et gen. rei” noted in lsj (9th ed.; 1996) s.v. ἀφίηµι A.ii.1.c, rather than the constructions used in early Christian literature (which will be discussed below).
46 There are several other meanings of ἀφίηµι mentioned in lsj (9th ed.; 1996) s.v. ἀφίηµι, which I will not analyze here.
47 Introductions to construction grammar can be found in e.g. J.-O. Östman and M. Fried, “Historical and Intellectual Background of Construction Grammar,” in Construction Grammar in a Cross-Language Perspective (ed. M. Fried & J.-O. Östman; Amsterdam: John Benjamins, 2004) 1-10; M. Fried and J.-O. Östman, “Thumbnail Sketch”; A.E. Goldberg, Constructions: A Construction Grammar Approach to Argument Structure (Chicago: University of Chicago Press, 1995); T. Hoffmann and G. Trousdale, “Construction Grammar: Introduction,” in The Oxford Handbook of Construction Grammar (Oxford: Oxford University Press, 2013) 1-14.
In New Testament scholarship, the use of construction grammar (and its sister theory, frame semantics) has been pioneered by Simon Wong and Paul Danove (references in note 4). Their usage of case grammar (one of the predecessors of construction grammar) has been criticized in S.E. Porter and A.W. Pitts, “New Testament Greek Language and Linguistics in Recent Research,” Currents in Biblical Research 6 (2008) 214-255, 228-230, since case grammar does not adopt a typology-based approach to semantic roles. However, for the purposes of this paper, that criticism is irrelevant, since I am primarily concerned with the pairing of syntax and semantics into constructions.
49 Construction grammar analyses can look very technical, but I will avoid unnecessary formalization. A full description of a construction should contain exhaustive information about it, for instance phonology, morphology, syntactic properties, evoked semantic frames and pragmatic information. However, most analyses that utilize construction grammar are selective and describe only those aspects that are relevant for a particular problem. M. Fried and J.-O. Östman, “Thumbnail Sketch.”
51 bdag s.v. ἀφίηµι 1 defines this usage similarly as “to dismiss or release someone or someth. from a place or one’s presence.” Cf. L&N §15.43-44.
52 On using construction grammar to structure dictionary entries, see Danove, Verbs of Transference, 168-170. On the difficulty of writing dictionary entries, see. J.A.L. Lee, A History of New Testament Lexicography (New York: Peter Lang, 2003) 3-13.
54 Against the possibility that σου is a genitive of separation and thus a beneficiary rather than a genitive attribute to αἱ ἁµαρτίαι speaks a) that no such usage of genitive with ἀφίηµι is noted in either lsj or bdag; b) the existence of “syntactic contamination,” which will be discussed below.
55 W.D. Davies and D.C. Allison, A Critical and Exegetical Commentary on the Gospel According to Saint Matthew, Vol. 2, Commentary on Matthew viii-xviii (icc; Edinburgh: T&T Clark, 1991) 89. Cf. Str-B 1.495-496.
57 Scholars do not agree whether “bind” and “loose” in Matt 18:18 refer to the right to issue halakhic prescripts (which is a reasonable interpretation of the parallel in 16:18) or the right to mediate divine forgiveness (or both). Given the literary context of 18:18, I agree with Ulrich Luz that the inclusion of the latter is probable for at least 18:18: U. Luz, Matthew 8-20: A Commentary (Minneapolis, Minn.: Fortress, 2001) 454-455.
58 R.H. Gundry, Matthew: A Commentary on His Handbook for a Mixed Church under Persecution (2nd ed.; Grand Rapids, Mich.: Eerdmans, 1994) 165.
59 J.L. Bybee, “Usage-based Theory and Exemplar Representations of Constructions,” in The Oxford Handbook of Construction Grammar (Oxford: Oxford University Press, 2013) 49-69.
62 A.C. Harris and L. Campbell, Historical Syntax in Cross-Linguistic Perspective (Cambridge: Cambridge University Press, 1995) 117-118.
63 C. Caragounis, New Testament Language and Exegesis: A Diachronic Approach (wunt I/323; Tübingen: Mohr Siebeck, 2014) 99-111.
64 J. Jeremias, New Testament Theology: Vol. 1, The Proclamation of Jesus (London: scm, 1971) 9-14; cf. Danove, Linguistics and Exegesis, 121-124.
65 T. Hägerland, Jesus and the Forgiveness of Sins: An Aspect of his Prophetic Mission (Cambridge: Cambridge University Press, 2011).
68 Pascut distinguishes between proper passivum divinum where God is the direct agent and indirect passivum divinum where God is the ultimate agent, but not the direct agent.
69 Talking about sin as debt is an Aramaism: J. Jeremias, The Prayers of Jesus (London: scm, 1967) 92. Cf. the discussion above about Anderson, Sin.
70 Unfortunately, space does not allow a detailed comparison with the Septuagint, but note that the few passages in the Septuagint that deal with interpersonal forgiveness and use the verb ἀφίηµι explicate the patient argument (Gen 50:17; Sir 28:2; 1 Macc 13:39; cf. Exod 10:17; 1 Sam 15:25; 25:28). The Septuagint also allows the patient in discourse on interpersonal forgiveness to be ἁµαρτία/ἁµάρτηµα (Gen 50:17; Exod 10:17; 1 Sam 15:25).
71 E.g. the article by J.R. Mantey in jbl arguing for the former position, refuted by Henry J. Cadbury in the same issue: J.R. Mantey, “The Mistranslation of the Perfect Tense in John 20:23, Mt 16:19, and Mt 18:18,” jbl 58 (1939) 243-249; H.J. Cadbury, “The Meaning of John 20:23, Matthew 16:19, and Matthew 18:18,” jbl 58 (1939) 251-254.
72 Cf. the scholarly discussion about whether God or a human community member is the agent in 1 John 5:16, δώσει αὐτῷ ζωήν. I have argued that the agent is a human intermediary of divine forgiveness in R. Roitto, “Practices of Confession, Intercession and Forgiveness in 1 John 1.9; 5.16,” nts 58 (2012) 232-253.
73 Mark 11:25 is more problematic to analyse than the other passages, since the phrase εἴ τις sometimes functions as an equivalent to the relative pronoun ὅστις in Hellenistic Greek (bdf §376). That is, εἴ τι ἔχετε κατά τινος, could be said to fulfil a semantic function which is equivalent to the patient argument. Should we argue that Mark 11:25a is a clause about interpersonal forgiveness with a patient argument but no beneficiary argument, similar to the exception in John 20:23 (discussed above)? I think not. Rather, as Mark wanted to clarify what should be forgiven in a clause about interpersonal forgiveness, it seems like his linguistic intuitions told him that he should not use ἀφίηµι with a direct object but rather find an alternative turn of phrase, and for that purpose the phrase εἴ τις does the trick.
74 For overviews of the discussion, see e.g. J. Delobel, “Lk 7,47 in Its Context: An Old Crux Revisited,” in The Four Gospels 1992: Festschrift for Frans Neirynck (ed. F. van Segbroeck et al.; Leuven: Leuven University Press, 1992) 1581-1590; Hägerland, Jesus, 51-59; J.J. Kilgallen, “Forgiveness of Sins (Luke 7:36-50),” NovT 40 (1998) 105-116.
- 28
Danove, Linguistics and Exegesis, 49-52, 160; Danove, Verbs of Transference, 13-17.
- 29
E.g. G. Helbig and W. Schenkel, Wörterbuch zur Valenz und Distribution deutscher Verben (8th ed.; Tübingen: Niemeyer, 1991) 31-39, use the term “fakultative Valenz.”
- 34
G. Lakoff and M. Johnson, Metaphors We Live By (Chicago: University of Chicago Press, 1980); G. Lakoff and M. Johnson, Philosophy in the Flesh: The Embodied Mind and Its Challenge to Western Thought (New York: Basic Books, 1999).
- 39
Cf. T. Kazen, Issues of Impurity in Early Judaism (Winona Lake: Eisenbrauns, 2010) esp. 13-40; J. Klawans, Impurity and Sin in Ancient Judaism (New York: Oxford up, 2000).
- 41
W. Bauer, Griechisch-deutsches Wörterbuch zu den Schriften des Neuen Testaments und der frühchristlichen Literatur (ed. K. Aland and B. Aland; 6th ed.; Berlin: de Gruyter, 1988) s.v. ἀφίηµι.
- 44
K.H. Jobes and M. Silva, Invitation to the Septuagint (Grand Rapids, Mich.: Baker Academic, 2000) 108-109; cf. M. Silva, Biblical Words and Their Meaning: An Introduction to Lexical Semantics (rev. and expanded ed.; Grand Rapids, Mich.: Zondervan, 1994) 53-100.
- 58
R.H. Gundry, Matthew: A Commentary on His Handbook for a Mixed Church under Persecution (2nd ed.; Grand Rapids, Mich.: Eerdmans, 1994) 165.
- 59
J.L. Bybee, “Usage-based Theory and Exemplar Representations of Constructions,” in The Oxford Handbook of Construction Grammar (Oxford: Oxford University Press, 2013) 49-69.
- 62
A.C. Harris and L. Campbell, Historical Syntax in Cross-Linguistic Perspective (Cambridge: Cambridge University Press, 1995) 117-118.
- 65
T. Hägerland, Jesus and the Forgiveness of Sins: An Aspect of his Prophetic Mission (Cambridge: Cambridge University Press, 2011).
