HEBREW MORPHOLOGICAL ANALYSIS by Alan Groves Westminster Theological Seminary Philadelphia, PA 19118 Last Revised 10/24/87 This file is organized into three parts: 1) An introduction 2) A quick overview of the system of analysis 3) A guide to the transliteration for entry and root 4) Supplements 10/14/87 1) Introduction The morphologically analyzed version of Ruth has been included on the CD ROM as an example of the form of the completed morphology. Grace of funding from the Packard Humanities Institute (PHI) and access to hardware and software provided by Dr. David Packard, the database is being corrected and verified at Westminster Seminary. We are working with an automatically analyzed version done on the IBYCUS (mini) by means of software written by Dr. Richard Whitaker (Pella, IA). The verification involves the comparison of three versions of the analysis: Whitaker's, one done at the Abbaye de Maredsous (Belgium) and a parsing of the verbs (only) done by Dr. Todd Beall, Capitol Seminary (Lanham, MD). Even-Shoshan's Hebrew Concordance is being used as the standard against which all corrections are being made. Where we depart from the root or analysis of Even- Shoshan, we flag the text and include a comment in a remark file to explain our reasoning. The date of the database's availability is uncertain. (Westminster Seminary also corrected the Michigan-Claremont version of the text of Biblia Hebraica Stuttgartensia which is elsewhere on this CD ROM.) The following is a sample of the first five entries from the sample text of Ruth as it now stands: 1:1 >Ru1:1 1:1 WA W@Pc 1:1 Y:HIY HYH@vqis3mj 1:1 BI B@Pp 1:1 YM"Y YOWM@ncpmc It will be changed somewhat in the final version. All shewa, dagesh and maqqef will be added (some were omitted for the purpose of the machine parsing done by Richard Whitaker.) The entry will also contain a marker to indicate accent position. One final difference is that we have added a number to indicate the sequential position of a word in a verse and a number to indicate the morpheme within the word that is being analyzed: >Ru1:1 1:1,1.1 WA W@Pc first word, first morpheme 1:1,1.2 Y:HI^Y HYH@vqis3mj first word, second morpheme 1:1,2.1 B.I B.@Pp second word, first morpheme 1:1,2.2 YM"Y^ YOWM@ncpmc second word, second morpheme (**Also added following the citation will be the flagging notation used in the machine-readable version of the running BHS text -- see elsewhere on the CD ROM for a sample of this text and an explanation of the notations. With words of multilple parts, each part will be marked with the flag even where only one part is actually affected.) ********************************************************************* ********************************************************************* SYSTEM OF ANALYSIS '@'separates the root from the analysis. What follows is a list of categories, determined by the symbol following the '@'. Upper and lower case is strictly observed. CATEGORY #1 Paragraphing CATEGORY #4 Noun @x = Paragraph Marker @ncg#s* = common noun (P@x or S@x only) @np1 = proper name (np2=second name of compound) CATEGORY #2 Particle CATEGORY #5 Adjective/Numeral @Pa = article @ag#s* = adjective @Pc = conjunction @Pd = adverb CATEGORY #6 Verb @Pg = interrogative @vsapg#s** = verb @Pi = interjection (Verbal roots=unvocalized) @Pn = negative @Pp = preposition CATEGORY #7 Suffixes @Pr = relative (Non-pronominal suffixes) @sh = paragogic Heh CATEGORY #3 Pronoun @sd = directional Heh @pipg#* = independent @sn = paragogic Nun @pspg#* = suffixed @pii = interrogative _________________________________________________________________ *For pronouns, common nouns and adjectives: (pg# = person, gender, number; g#s = gender, number, state) p g # s 1=first person m=masculine s=singular c=construct 2=second person f=feminine p=plural 'blank'= 3=third person d=dual absolute _________________________________________________________________ **For Verbs: (sapg#s=stem,aspect,person,gender,number,state -- for pg#s, see above) s a other q=qal p=perfect j=apocopated form n=nifal i=imperfect h=cohortative form p=piel v=imperative P=pual c=infinitive construct t=hitpael a=infinitive absolute h=hifil P=particple H=hofal s=qal passive participle Q=truly qal passive Polel type stems are treated as regular piel, pual, hitpael. ********************************************************************* ********************************************************************* See CODES.TXT for the Hebrew transliteration key. ********************************************************************* ********************************************************************* SUPPLEMENT TO THE CODE MANUAL FOR THE MICHIGAN OLD TESTAMENT by Alan Groves Westminster Theological Seminary Philadelphia, PA 19118 Last Revised 10/14/87 1. Introduction 1.1 This document is a supplement to "Code Manual for the Michigan Old Testament". Knowledge of that document is presupposed at all points which follow. "MCMOT" will be used to indicate references to the Michigan manual in the following material. 1.2 The present machine-readable text of the Michigan-Claremont version of the text of BHS has been verified by means of a computer aided comparison with an the following machine- readable texts: 1.2.1 An independently encoded text from the Abbaye de Maredsous 1.2.2 An independently proofread version of the M-C text from Bar Ilan University. 1.3 Moreover, verification of certain features (e.g. position of cantillation within each word) was done automatically by means of software written for that purpose. 1.4 For suspected problems with the printed text of BHS (1983 edition), we appealed to the two other published versions of the Codex Leningradensis: 1.4.1 "The Holy Scriptures" edited by Aron Dotan (Adi, Tel Aviv, 1974) 1.4.2 BHK (Third edition) 1.5 At all points of variance between Dotan and BHS or BHS and BHK, we also examined the photo facsimiles of the codex (Codex Leningradensis b19A, D.S. Loewinger, Makor, Jerusalem, 1971.) 1.6 Two notes are important for further understanding our work: 1.6.1 The United Bible Society did not work from the Makor photo-facsimiles in producing BHS. Rather, they worked from a separate set of microfilm. Having now seen both photographic reproductions, we are convinced that Makor is clearer at almost all points of dispute. Certain features of the codex that are invisible on the microfilm show up in the Makor photographs. We hope to publish an article concerning this. 1.6.2 Because some typographical errors in the 1977 edition have already been corrected in the 1983 edition, we used the latter as the basis of our proofreading. (The 1977 edition was the basis for the input). As a result, several of the ']' flags that were in earlier versions of the machine-readable text have been removed. 1.7 The results of our work therefore, are as follows: 1.7.1 A text that, we believe, improves on BHS in terms of reflecting the codex (Makor photofacsimile) 1.7.2 A text which is more accurate than other currently available machine-readable texts. 2. The current version of the text differs from the description given in the MCMOT as follows: 2.1 The logical record length is now variable (up to 80 characters). See MCMOT 2.1.2. 2.2 Unlike earlier versions of the database, each verse has the chapter number listed in its citation. (This is an exception to MCMOT 2.1.4.) 2.2.1 For example, Genesis 1:2 will begin with '1:2' not '2'. 2.3 The second and third records of each file have been removed. These records listed the various ASCII characters used in the transliteration of the text. (This is an exception to MCMOT 2.1.3.) 2.4 The Ketib-Qere is as described in MCMOT 3.6 with the following exceptions: 2.4.1 Contrary to MCMOT 3.6.2, KQ involving only the first word of a compound joined by a maqqef will be done differently than is indicated in the manual. (The second example in MCMOT 3.6.2 is the issue here.) The maqqef always belongs to the Ketib, never the Qere. This will facilitate printing. 2.4.1.1 Ruth 4:6 becomes *LIG:)FWL-LI80Y **LIG:)FL 2.4.1.2 Also, note that Exodus 21:8 is ):A$ER-?*LO) (See Supplement 2.6 on '?') 2.4.2 The Ketib wela Qere is indicated by the unvocalized Ketib followed by **zz. The Qere wela Ketib is *zz followed by the Qere. (This represents a change to the statements at the end of the first paragraph of MCMOT 3.6.) 2.5 MCMOT 4.2 is no longer accurate. ']', and not '!', is used to mark deviations from BHS or to note other special features. In particular, we have expanded the coding to make the flagging procedure more precise by adding a single character or numeral following the bracket to indicate something special about the text. (For all the details, see the end of this supplement document entitled "Explanation of the Right-Hand Bracket".) 2.5.1 The user should remember that we based our corrections on the 1983 edition of BHS. (See Supplement 1.6.2.) We do not retain a bracket where the 1983 edition has already made a correction to the 1977 edition. 2.6 The cantillation coding varies from that which is described in MCMOT 3.5 as follows: 2.6.1 Referring to the Tabula Accentuum which accompanies BHS, #10 in the Accentus Communes lists a second case for the occurrence of the pasta in tandem. Originally, the first mark was encoded as an 'azla ('63' in the Michigan format) and the second as a pasta ('03' in the Michigan format). We have made the first mark a '33'. Note that it is written after the letter not over it (unlike the 'azla). The coding of the pasta remains '03'. 2.7 A '?' is used to mark the physical end of the line as it appears in BHS. 2.7.1 A space will follow a '?' except where a maqqef ('-') ends a line in BHS. (e.g. Gn 1:2 W:HF)F81REC)? HFY:TF71H) 2.7.2 In that case, the '-?' will be followed by the next word without intervening space. (e.g. Gn 1:4 )ET-?HF)O71WR) 2.7.3 The '?' has not been verified systematically at this point. 2.8 We have adapted the premorphological encoding ('/') to reflect the citations in Even Shoshan's concordance. (We recognize that the concordance is based on the Koren edition of the Bible. However, the differences between the Koren text and BHS have no impact on this issue.) 2.8.1 One result is that there is a standard against which our work can be evaluated and that there is only one standard used. 2.8.2 This means that any supposed compound form that is listed as a separate entitiy under its compound form (as opposed to being listed solely under one or both components of the compound) is treated as a unit. In such cases no premorphological divider has been used. Listed below are the 6 exceptions to this rule. In each case the premorphological divider has been kept to indicate a prefix (a caret marks the place where the accent would most likely occur). 2.8.2.1 K.F/M.F^H 2.8.2.2 LF/K"^N 2.8.2.3 L:/PI^Y 2.8.2.4 LF^/M.FH 2.8.2.5 LI/P:N"^Y 2.8.2.6 MI/P:N"^Y 3. SPECIAL NOTE #1 The following items have not been verified as thoroughly as other items in the text: 3.1 The premorphological indicator '/' will be more comprehensively verified in the morphological verification yet to come. 3.2 The '?' which marks the physical end of the line in BHS has not been verified consistently. 3.3 The accuracy of the right-hand and left-hand cantillation markings require checking against the codex (i.e. 75 and 95 in the text.) 4. SPECIAL NOTE #2 No matter how careful we have been, we have certainly left some errors. For example, any time that the encoding of Maredsous text was wrong in the same way as the M-C encoding, we are not likely to have picked up the error. (The probability of making the same error at the same place is not high. We know of only one case. The one case was uncovered by a scribe doing spot checks on a printed version of our proofread text (printed in Hebrew, fully vocalized and cantillated). In this way about 12% of the text was checked and only this one coincidence of error was observed.) While we believe remaining errors are minimal, some still exist. *********************************************************** *********************************************************** EXPLANATION OF THE RIGHT-HAND BRACKET ']*' (All bracketing has been done on the basis of the 1983 edition, not the 1977 edition. Moreover, the Makor edition of the codex was checked at all relevant points in determining the need for a flag.) ]1 BHS has been faithful to the Leningrad Codex where there might be a question of the validity of the form and we keep the same form as BHS. e.g. Deuteronomy 23:18 YI&RF)"L00]1 (missing silluq) ]2 We have added a sop pasuq where L and BHS omit it. ********** (Special note: Formerly this category had a much broader range and read as follows: "BHS has been faithful to the Codex where there might be a question of the validity of the form and we have abandoned BHS in order to code it differently." All situations other than missing sop pasuqs have now been made to conform to L and then labeled as ]1. ********** ]3 We read or understand L differently than BHS (1983 Edition). Often this notation indicates a typographical error in BHS. e.g. Genesis 6:22 ):ELOHI73YM]3 ]4 Puncta Extraordaria -- a 52 is used to mark such marks in the text when they are above the line and 53 when they are below the line. e.g. Genesis 18:9 )"52LF8052Y52W52]4 ]5 Large letter(s) ]6 Small letter(s) ]7 Suspended letter(s) ]8 Inverted Nun (N]8 in the text) ]9 BHS has abandoned L and we concur. All of these occurrences are ketib/qere problems. ]q We have abandoned or added a ketib/qere relative to BHS. In doing this we agree with L against BHS. e.g. Exodus 32:17 B.:/R:(O92H]q ]a Adaptations to a Qere which L and BHS, by their design, do not indicate. e.g. Exodus 4:2 **-Z.E74H]a ]y Yathir readings in L which we have designated as Qeres when both Dothan and BHS list a Qere. ]m Miscellaneous notes to the text and occasions where more than one bracket category applies.