Anaphora Resolution In a Biblical Passage Final Draft

Anaphora Resolution In a Biblical Passage 1

Anaphora Resolution in a Biblical Passage Eduard C. Hanganu B.A., M.A., Linguistics Lecturer in English, UE

Draft 10 Revised – December 19, 2014 © 2014

Anaphora Resolution In a Biblical Passage 2

Anaphora Resolution in a Biblical Passage Introduction Daniel, the presumed “encoded” Bible book, has been for millennia a persistent puzzle for Hebrew scholars and theologians, but a traditional grammatical approach that has attempted to “decode” the text has not been too successful. There are still, in Daniel’s chapters, countless interpretation issues that await clarification, and frustrated theologians and scholars seem to have hit the traditional brick wall in their efforts to find solutions to such problems. One of the most vexing enigmas is the origin of the little horn in Daniel 8: 8. The matter has been studied, deliberated, and argued without too much success in various theological circles for quite a long time, and the issue still remains to be settled. Some biblical scholars contend that the little horn arose from the “four notable” horns that followed the dissolution of Alexander the Great’s immense empire, while other scholars and theologians are certain that the little horn came out of one of the “four winds of heaven” mentioned in the same biblical text. This paper argues that although the topic is theological in nature and the efforts to solve the little horn riddle have been for the most part dogmatic—deductive and non-empirical rather than inductive and empirical—there is a better approach to the problem’s solution, and that approach is based in a linguistic method—anaphora resolution. Evidence provided in this paper will show that the anaphora resolution approach can help solve the puzzle that surrounds the little horn in Daniel 8, and provide a logical and factual clarification for the biblical text. The Biblical Text Under Discussion The biblical passage that has puzzled numerous theologians and Bible scholars for centuries is located in Daniel 8: 8-9, and the question that still waits for an answer is what would

Anaphora Resolution In a Biblical Passage 3

be the correct grammatical antecedent to the pronoun “them” in verse 9. The English translation below comes from the King James Bible: KJV 8

Therefore the he goat waxed very great: and when he was strong, the great horn was

broken; and for it came up four notable ones [emphasis added] toward the four winds of heaven [emphasis added]. 9

And out of one of them [emphasis added] came forth a little horn, which waxed

exceeding great, toward the south, and toward the east, and toward the pleasant land.

The Confusion About Gender Discord The complication that intervenes in the traditional, that is, grammatical, resolution process for the antecedent to the pronoun “them” in Daniel 8:9 is a gender discord that occurs between verses 8 and 9 in Daniel 8. One theologian explains this Hebrew language gender discord in the following comment: 9. Out of one of them. In the Hebrew this phrase presents confusion of gender. The word for “them,” hem, is masculine. This indicates that, grammatically, the antecedent is “winds” (v. 8) and not “horns,” since “winds” may be either masculine or feminine, but “horns,” only feminine. On the other hand the word for “one,” achath, is feminine, suggesting “horns” as the antecedent. Achath could, of course, refer back to the word for “winds,” which occurs most frequently in the feminine. But it is doubtful that the writer would assign two different genders to the same noun in such close contextual relationship. To reach grammatical agreement, either achath should be changed into a masculine, thus making the entire phrase refer clearly to “winds,” or the word for “them”

Anaphora Resolution In a Biblical Passage 4

should be changed into a feminine, in which case the reference would be ambiguous, since either “winds” or “horns” may be the antecedent. A number of Hebrew manuscripts have the word for “them” in the feminine. If these manuscripts reflect the correct reading, the passage is still ambiguous. (Nichol, 1978) The “Nearest Antecedent” Explanation Because Daniel 8:8 appears to provide two choices (1. “four notable ones, and 2. “four winds of heaven”) for the pronoun “them,” some theologians and scholars have claimed that the solution to the problem might be found in the “nearest antecedent” notion, that is, that the “nearest antecedent” to the pronoun “them” in Daniel 8:9 might be the true referent for the pronoun “them” because “the ‘them’ in ‘out of one of them’ at the beginning of verse 9 most naturally refers to the nearest antecedent [emphasis added]: the immediate preceding ‘four winds of heaven’ at the end of verse 8” (Gane, 2006, p. 37). Gane’s proposed solution, though it seems to be based in “common sense” and deductive logic does not stand the empirical test. It appears that the issue requires more than common sense and deduction—it requires induction that ends in a conclusion based on empirical or factual evidence. In the conclusion to his research about the possible statistical distance between the anaphor and its true antecedent, Mitkov (1999) had warned: Most of the anaphora resolution systems deal with resolution of anaphors which have noun phrases as their antecedents because identifying anaphors which have verb phrases, clauses, sentences or even paragraphs/discourse segments as antecedents, is a more complicated task. Typically, all noun phrases (NPs) preceding an anaphor are initially regarded as potential candidates for antecedents. Usually, a search scope has to be

Anaphora Resolution In a Biblical Passage 5

identified: most approaches look for NPs in the current and preceding sentence. However, an "ideal" anaphora resolution system should extend its scope of search: antecedents which are 17 sentences away from the anaphor have already been reported [emphasis added]. (p.3)

Gender Discord and Diglossia Puzzle The gender discord in between verses 8 and 9 in Daniel 8 could be another deductive trap when conclusions are based on opinions and not on research. The grammatical discord between the two verses is a natural language phenomenon in Hebrew—diglossia—that happens when two different Hebrew language varieties are in use and has no weight in the anaphora resolution process for the pronoun “them.” Explains Laiu: The gender disagreement in Da 8:9 and consequently the grammatical chiastic agreement suggested50 by scholars have been best explained by Martin Pröbstle,51 following Rendsburg’s studies.52 While this masculine form ‫ ֵמהֶם‬mē·hém, instead of the required feminine form ‫ ֵמהֶן‬mē·hén is not Standard Biblical Hebrew, and would be considered as a grammatical disagreement, it is merely a formal disagreement, specific to the spoken Hebrew that often uses masculine forms for both genders.53 (2013, pp. 19-21) Anaphora Resolution And Its Method The scientific and reliable approach for the solution to the anaphora problem in Daniel 8:8-9 that linguistics provides, anaphora resolution solves the issue and eliminates the confusion about what the authentic antecedent to “them” might be. Orăsan and Evans define anaphora resolution as follows:

Anaphora Resolution In a Biblical Passage 6

Anaphora resolution is the process which attempts to determine the meaning of expressions such as pronouns or definite descriptions whose interpretation depends on previously mentioned entities or discourse segments [emphasis added]. Anaphora resolution is very important in many fields of computational linguistics such as machine translation, natural language understanding, information extraction and text generation (Mitkov, 2002). (2007, p. 79)

Prolo delivers a short but clear explanation about the natural discourse references introduced with previous expressions, the distinction between anaphor and its antecedent and the discourse relationship between the two, and then restates in more explicit and detailed terms what the antecedent or pronoun resolution approach involves: Anaphora is roughly speaking1 the phenomenon of making in a discourse abbreviated references to entities that have been directly or indirectly introduced by a previous expression. The expression used to make the abbreviated reference is called the anaphor and the previous expression the antecedent. In this paper I restrict myself to the cases where the anaphor is a pronoun. The concept is illustrated in (1). The occurrence of the pronoun he makes an anaphoric reference to the discourse entity introduced by the noun phrase Carlos. Hence he is the anaphor and Carlos is the antecedent. There are two occurrences of the possessive pronoun its, the first has as antecedent the paper, and the second, as well as the occurrence of it, refers to the introduction of the paper. (1) Carlos is writing a paper. Right now he is rewriting its introduction because he is not happy with its current form. It should be clear enough so that people not familiar with the area can understand at least the topic the paper is about.

Anaphora Resolution In a Biblical Passage 7

By pronoun resolution I mean the process of finding for each pronoun its antecedent, and the rules that govern the choice of the antecedent by the hearer/listener are the central point of this problem that the proposals here discussed try to capture [emphasis added]. (2006, 139-140) The anaphora or pronoun resolution process involves work with multiple and verified linguistic “constraints” and “preferences” that will allow the scholar to distinguish the true and genuine noun antecedent from other possible but illegitimate antecedents and resolve the anaphora or pronoun antecedent issue. Mitkov, professor of Computational Linguistics and anaphora resolution expert at the British University of Wolverhampton, describes the anaphora resolution process as follows:

1.3 The process of anaphora resolution Most of the anaphora resolution systems deal with resolution of anaphors which have noun phrases as their antecedents because identifying anaphors which have verb phrases, clauses, sentences or even paragraphs/discourse segments as antecedents, is a more complicated task. Typically, all noun phrases (NPs) preceding an anaphor are initially regarded as potential candidates for antecedents. Usually, a search scope has to be identified: most approaches look for NPs in the current and preceding sentence. However, an “ideal” anaphora resolution system should extend its scope of search: antecedents which are 17 sentences away from the anaphor have already been reported [emphasis added] (Mitkov 1995a)! Assuming that the scope of search for a specific approach has already been specified, the NPs preceding the anaphor within that scope are identified as candidates for antecedents

Anaphora Resolution In a Biblical Passage 8

and a number of anaphora resolution factors are employed to track down the correct antecedent. Approaches to anaphora resolution usually rely on a set of “anaphora resolution factors.” Factors used frequently in the resolution process include gender and number agreement, c-command

constraints,

semantic

consistency,

syntactic

parallelism,

semantic

parallelism, salience, proximity etc. These factors can be “eliminating” i.e. discounting certain noun phrases from the set of possible candidates (such as gender and number constraints3, c-command constraints, semantic consistency) or “preferential,” giving more preference to certain candidates and less to others (such as parallelism, salience). Computational linguistics literature uses diverse terminology for these – for example E. Rich and S. LuperFoy (Rich & LuperFoy 1988) refer to the "eliminating" factors as “constraints,” and to the preferential ones as “proposers,” whereas Carbonell and Brown (Carbonell & Brown 1988) use the terms “constraints” and “preferences.” Other authors argue that all factors should be regarded as preferential, giving higher preference to more restrictive factors and lower – to less “absolute” ones, calling them simply “factors” (Preuß et al. 1994), “attributes” (Rico Pérez 1994), "symptoms" (Mitkov 1995b) or “indicators” (Mitkov 1996a, 1998b). The division of factors into constraints and preferences has led to distinguishing between constraint-based and preferences-based architectures in anaphora resolution (Mitkov 1997b). 1.3.1 Constraints Several constraints will be outlined and illustrated by examples. Coreferential items are given the same index.

Anaphora Resolution In a Biblical Passage 9

• Gender and number agreement This constraint requires that anaphors and their antecedents must4 agree in number and gender. Example: Janei told Philipk and his friendsm that shei was in love. • Syntactic binding theories’ constraints Results in Government and Binding Theory (GB)

5

and Lexical Functional Grammar

have provided useful constraints on the anaphors and their antecedents which have been successfully used in anaphor resolution. For instance, various GB c-command restrictions have been formulated in (Ingria & Stallard 1989) for eliminating unacceptable candidates when searching for the antecedent: (a) A non-pronominal NP cannot overlap in reference with any NP that c-commands it. Hei told them about Johnj. (b) The antecedent of a bound anaphor must c-command it. Johni likes pictures of himselfi. (c) A personal pronoun cannot overlap in reference with an NP that c-commands it. Johni told Billj about himk. • Semantic consistency This constraint stipulates that if satisfied by the anaphor, semantic consistency constraints must be satisfied also by its antecedent. Vincent removed the diskette from the computeri and then disconnected iti. Vincent removed the diskettei from the computer and then copied iti.

Anaphora Resolution In a Biblical Passage 10

1.3.2 Preferences Preferences, as opposed to constraints, are not obligatory conditions and therefore do not always hold. We shall illustrate three preferences: syntactic parallelism, semantic parallelism and center of attention. • Syntactic parallelism Syntactic parallelism could be quite helpful when other constraints or preferences are not in a position to propose an unambiguous antecedent. This preference is given to NPs with the same syntactic function as the anaphor. The programmeri successfully combined Prologj with C, but hei had combined itj with Pascal last time. The programmeri successfully combined Prolog with Cj, but hei had combined Pascal with itj last time. • Semantic parallelism This is a useful (and stronger than syntactic parallelism) preference but only systems which can automatically identify semantic roles, can employ it. It says that NPs which have the same semantic role as the anaphor, are favoured. Vincent gave the diskette to Sodyi. Kim also gave himi a letter. Vincenti gave the diskette to Sody. Hei also gave Kim a letter. • Centering Although the syntactic and semantic criteria for the selection of an antecedent are very strong, they are not always sufficient to distinguish between a set of possible candidates. Moreover, they serve more as filters to eliminate unsuitable candidates than as proposers of the most likely candidate. In the case of antecedent ambiguity, it is the most salient

Anaphora Resolution In a Biblical Passage 11

element among the candidates for antecedent which is usually the frontrunner. This most salient element is referred to in computational linguistics as focus (e.g. (Sidner 1979) [sic!] or center6 e.g. (Grosz et al. 83) though the terminology can be much more diverse (Hirst 1981; Mitkov 1995a). For instance, neither machines, nor humans, would be able to resolve the anaphoric pronoun “it” in the sentence Jenny put the cup on the plate and broke it. However, if this sentence is part of a discourse segment7 which makes it possible to determine the most salient element, the situation is different: Jenny went window shopping yesterday and spotted a nice cup. She wanted to buy it, but she had no money with her. Nevertheless, she knew she would be shopping the following day, so she would be able to buy the cup then. The following day, she went to the shop and bought the coveted cup. However, once back home and in her kitchen, she put the cup on a plate and broke it... In this discourse segment, “the cup” is the most salient entity and is the center of attention throughout the discourse segment. It is now clear that very often when two or more candidates “compete” for the antecedent, the task of resolving the anaphor is shifted to the task of tracking down the center/focus of the sentence (clause). Various methods have already been proposed to center/focus tracking (e.g. Brennan et al. 1987; Dahl & Ball 1990; Mitkov 1994b; Sidner 1986; Stys & Zemke 1995; Walker et al. 92). However useful the term center (or focus) can be for anaphora resolution, we should point out that it has suffered from two inconveniences: its intuitive nature and the use of

Anaphora Resolution In a Biblical Passage 12

different terms to describe concepts which either seem to be very close to “center” or even could be considered practically identical (e.g. focus, topic, theme - for further details please see (Hirst 1981) and (Mitkov 1995a). 1999, pp. 3-6) The Anaphora Resolution Solution Table Based on the constraints and preferences data taken from Mitkov (1999, pp. 3-6, and 1997, pp. 14-17) and other world experts in anaphora resolution (Brasoveanu, 2010; Kehler, 1997 pp. 467-475; Brennan, Walker Friedman, & Pollard, 1987, pp.155-162; Struckardt, pp. 1-6; Orasan & Evans, 2007, pp. 79-103) the above-mentioned selection parameters have been organized into a table that will be used to solve the pronoun resolution issue in Daniel 8:8-9. The biblical texts that will be considered in the anaphora or pronoun resolution process are included again below: KJV8

Therefore the he goat waxed very great: and when he was strong, the great horn was broken; and for it came up four notable ones toward the four winds of heaven [emphasis added]. KJV9

And out of one of them [emphasis added] came forth a little horn, which waxed exceeding great, toward the south, and toward the east, and toward the pleasant land. The various constraints and preferences that will allow the distinction and separation between the true and legitimate noun antecedents and other possible but illegitimate antecedents in the anaphora or pronoun resolution process have been organized into the table below that will facilitate the resolution process and provide an overall perspective on the linguistic approach: Anaphora Resolution in Daniel 8:8-9 Resolution Parameters

Constraints

Little Horn From “Them” (“Notable Ones [Horns]”)

Little Horn From “Them” (“Winds of Heaven”)

Linguistic Data on the Parameters The constraints are used to eliminate certain noun

Anaphora Resolution In a Biblical Passage 13

+ Number agreement.

+ Number agreement.

+ Gender neutralization.

+ Gender agreement.

3. Person Agreement

+ Person agreement.

+ Person agreement.

4. Case Agreement

+ Construct relation.

+ Construct relation.

+ Outside local domain. + Direction mentioned. + Antecedent consistent.

+ Outside local domain. Direction not mentioned.

1. Number Agreement

2. Gender Agreement

5. Syntactic Constraints 6. Selectional Restrictions 7. Semantic consistency

8. Recency

9. Grammatical Role

a. Subject Position

b. Object Position

+ Both entities recent. + “Horns” more salient. + Subject position.

Subject position.

_

phrases (NPs) from the set of possible antecedent candidates. This constraint requires that the anaphor and its antecedent should agree in number. This constraint requires that the anaphor and its antecedent should agree in gender. This constraint requires that the anaphor and its antecedent should agree in person. This constraint requires that the anaphor and its antecedent should agree in case. The pronouns choose their antecedents outside of their local domain. A verb places restrictions on its arguments.

If satisfied by the anaphor, Antecedent semantic consistency inconsistent. constraints must be satisfied also by its antecedent. + Entities introduced recently Both entities recent. are more salient than those introduced before. Entities mentioned in “Winds” less salient. subject position are more salient than those in object position. From the list of potential Object position. candidates, the subject of the previous sentence (clause) is the preferred antecedent. From list of potential + candidates, the subject of Object position. the previous sentence (clause) is the preferred

Anaphora Resolution In a Biblical Passage 14

c. Genitive Position

+ Construct relation.

+ Construct relation.

+ Subject – Subject.

Subject – Object.

+ Parallelism present.

Parallelism absent.

+ No semantic emphasis.

+ No semantic emphasis.

+ “Horns” are repeated. + Factual agreement. Furthest possible antecedent.

“Winds” not repeated. Factual disagreement. + Nearest possible antecedent.

+ “Horns” more sentient. + The horns are the discourse center.

“Winds” less sentient. The winds are not the discourse center.

Preferences

1. Syntactic Parallelism

2. Semantic Parallelism

3. Verb Semantics

4. Repeated Mention

5. World Knowledge

6. Distance

7. Animate

Discourse Center

antecedent; the second preferred antecedent is the direct object. From the list of potential candidates the subject of the previous sentence (clause) is the preferred antecedent; the second preferred antecedent is the direct object. The third preferred antecedent is in the genitive position. While the constraints rule out the implausible candidates, the preferences emphasize the proper selection of the most likely antecedent. This preference is given to the entities with the same syntactic function as the anaphor. The favored antecedents are those that have the same semantic role as the anaphor. Certain verbs appear to place a semanticallyoriented emphasis on one of their argument positions. Entities that have been focused on in the prior discourse are more salient. A horn cannot grow greater than the animal on which it sits. Candidates from the previous clause or sentence are preferred. Animate (sentient) entities are more salient that inanimate entities. Certain discourse entities are more central than the

Anaphora Resolution In a Biblical Passage 15

others.

Conclusion The anaphora resolution table submitted above shows that the two possible antecedents to “them” – the “notable ones [horns],” and the “winds of heaven” – obtain similar values for number, gender, person, case, syntactic constraints, and recency, but that these two potential antecedents differ concerning their grammatical roles (“horns” is the sentence subject, while “winds” is the sentence object), syntactic parallelism (the term “horn” is parallel to “horn,” while the term “horn” is not parallel to “wind”), repeated mention (“horn” is repeated in Daniel 8:8-9, while “wind” is never repeated in the texts), world knowledge (a “horn” cannot grow greater than the animal on which it sits), animation (horns are animated, while winds are not), and discourse center (“horn” is the discourse center, while “wind” is an adverbial and therefore peripheral). For these multiple reasons it is obvious that “wind” (verse 8) cannot be the legitimate and actual antecedent for the little “horn” (verse 9) and that the preferred and acceptable antecedent must be the “notable horn.”

Anaphora Resolution In a Biblical Passage 16

References Adrian Brasoveanu. (Spring 2010). Anaphora Resolution. Retrieved December 22, 2013 from people.ucsc.edu/~abrsvn/anaphora%20resolution.ppt Andrew Kehler. (1997). Current Theories of Centering for Pronoun Interpretation: A Critical Evaluation. Association for Computational Linguistics. Carlos Augusto Prolo. Computational Approaches to Pronoun Resolution. Letras de Hoje. Porto Alegre. v. 41, nº 2, Junho, 2006. Constantin Orăsan & Richard Evans. (2007). “NP Animacy Identification For Anaphora Resolution,” Journal of Artificial Intelligence Research. Florin Laiu. (2013). The Sanctuary Doctrine: A Critical-Apologetic Approach (Bucharest: Romania, 2013). Unpublished manuscript. Retrieved from http://www.academia.edu/4145197. Francis D. Nichol.(1978). The Seventh-day Adventist Bible Commentary. Washington, DC: Review and Herald Publishing Association. Roland Stuckardt, “Anaphor Resolution and the Scope of Syntactic Constraints,” German National Research Center for Information Technology (GMD) Retrieved on December 22, 2013 from acl.ldc.upenn.edu/C/C96/C96-2158.pdf, 1-6. Roy Gane. (2006). Who’s Afraid of the Judgment? Nampa, Idaho: Pacific Press Publishing Association. Ruslan Mitkov (1997). “Factors In Anaphora Resolution: They Are Not The Only Things That Matter. A Case Study Based On Two Different Approaches.” In Proceedings of the ACL'97/EACL'97 workshop on Operational factors in practical, robust anaphora resolution, Madrid, Spain.

Anaphora Resolution In a Biblical Passage 17

Ruslan Mitkov. (1999). Anaphora Resolution: The State Of The Art, working paper based on the COLING'98/ACL'98 tutorial on anaphora resolution (University of Wolverhampton, Wolverhampton. Susan E. Brennan, Marilyn Walker Friedman, and Carl J. Pollard (1987), A Centering Approach to Pronouns, 25th Annual Meeting of the Association of Computational Linguistics.