The text of Luke 1:38 is widely interpreted essentially as in the NRSV. No textual variations or grammatical issues of the Greek text are significant. The Greek word doule "female slave" in this text is commonly translated by the English term "servant", implying a legal status that scarcely existed in the Hellenistic world. This follows a West European tradition that apparently stems from the early Latin translations of the gospels. This translation is not appropriate in view of the knowledge of Hellenistic Greek and its legal terminology that has accumulated since major studies of the language and contents of the the Greek papyri from Ptolemaic Egypt began to appear about a century ago. Revision of the translation of this passage leads to a different understanding of Mary's response to the annunciation. Mary recognizes her subordinate position and what follows from it, no less, no more.
The NRSV text here reads, reporting the supposed words of Mary speaking in response to the words of the angel Gabriel that she will bear a child who will be called the Son of God:
"Here am I, the servant of the Lord; let it be with me according to your word".
The recent New Interpreters' Study Bible (Nashville 2003) notes "although introduced without reference to her family origins (v.27), Mary now claims a place in God's household. Partnership with God transcends family ties ( . . . )." This translation and note set the stage for the issue of the interpretation of this passage.
This is a passage where there are no significant variants of the Greek which might underly different understandings. The text is not attested in the papyri of the first three centuries AD, and the standard editions of the Greek text of Nestle/Aland (27th edition, 1999) and the United Bible Society (4th edition, Stuttgart, 1994) list only one minor variant. Other very minor variants of the text which need not be considered here are provided by Tischendorf (editio octava critica major, I, Leipzig 1869).
The Greek word doulos which is translated "servant" in the NRSV and in most English-language versions is more accurately to be translated "slave". Apparently on the basis of the etymological formation of the Greek word doulos, which contains the root "tie", the term "bondservant" or "bondswoman" has been used to translate it. Frederick Danker's new edition (Chicago, 2000) of Walter Bauer's wonderful standard dictionary of NT Greek defines the term clearly as "slave as an entity in a socioeconomic context" or "one who is solely committed to another, slave, subject". The word doulos is used in its concrete legal sense in references to legal slaves and is also used metaphorically in the second sense by speakers to those of rank much higher than theirs, such as officials, nobles, or God. This interpretation is based on the Bauer/Danker article on the word in the masculine gender (with literature); the much briefer article on the word in the feminine gender, where there are many fewer passages to consider, defines it as "female slave, bondswoman". Danker interprets the present passage as "an oriental expression, used by one of humble station in addressing one of higher rank or a deity". (p.259). Thus an interpretation such as that in the New Interpreter's Study Bible above, that Mary claims by its use a place as servant in God's household, is misleading.
The Bauer/Danker interpretation is not new. The distinguished German theologian Adolf Deissmann while at the University of Heidelberg wrote an epoch-making treatise Light from the Ancient East (1908, English 1910) in which he showed the importance of the late Greek of the Egyptian papyri for the interpretation of the Greek Bible and especially for the NT. Here Deissmann in a long excursus (pp.322 ff.) points out the significance for New Testament theology of the accurate understanding of the term doulos, which occurs in the feminine form in this passage. He discusses the use of the term especially in regard to Hellenistic popular law and references to manumission (in Greek, the same word as "redemption") in the NT, which lose their point if doulos is translated as "servant" rather than "slave". The point was also made by a British scholar who in terms of linguistic expertise in Hellenistic Greek usage was at least the equal of Deissmann, James Hope Moulton, in his Vocabulary of the New Testament(1914/1930). Moulton and his collaborator Milligan cite the study by the papyrologist Wilcken (1899) which points out the extent to which various occupations in the Hellenistic world were held exclusively either by slave labor or free labor. The issue of occupation thus overlaps the issue of legal status. It follows that, in the words of the International Critical Commentary volume on Luke by Alfred Plummer, "In an age in which almost all servants were slaves, the idea which is represented by our "servant" could scarcely arise" (Edinburgh 1901, p.26). The term used in the Greek NT defines Mary metaphorically in terms of her legal status (slave), not metaphorically in terms of her occupation. Since not all slaves in the Hellenistic world were house-servants, however, from Luke's Greek term "slave" it could not be inferred that the person to whom it referred was a servant.
However, the Latin translation which has overshadowed the Greek text in a tradition which we know from Jerome and is suggested even in the notes cited above translates the Greek term by a Latin one which often refers to occupation: "ecce ancilla Domini fiat mihi secundum verbum tuum" in the Biblia Sacra Latin text of the Wuerttembergische Bibelanstalt (ed. Robert Weber, Stuttgart 1969), which lists no textual variants. In Latin, the term "ancilla" is translated in the older (and broader) dictionary of Lewis and Short (OUP 1879) as "maidservant, handmaid, female slave" with the observation that it was commonly used as the feminine of servus. The newer Oxford Latin Dictionary ed. P.Glare (OUP 1996) treats the word more from a legal point of view ("female slave, slave girl, maidservant"), but Lewis and Short make clear the usage as a designation of occupation in later Latin than is treated in Glare's Oxford dictionary of classical Latin. Deissmann (p.323) points out that the German Bible translation of Luther also softens the Greek term "slave" and creates legal confusion, just as does the English translation by translating the Greek word "slave" by the English word "servant".
These traditions in West European Bible translations seem to reflect the influence of the early anonymous translators of the Greek NT into Latin whose work Jerome revised in his fourth-century Latin version of the gospels. The Latin text hovered in the back of the mind of early West European translators, even those working directly from the Greek, and its influence is not entirely absent even from modern scholarship. By contrast, this influence is lacking in the earliest East European translation, the Old Bulgarian translation of the NT of the mid-ninth century, which was carried out directly from the Greek, probably by Greeks who were bilingual in Old Bulgarian. There, the earliest surviving Old Bulgarian manuscripts of the gospels without an exception (I checked the six editions of them on my shelves) read the equivalent of "Behold the slave of the Lord" in this verse, which is also a correct rendering of the Greek original.
The second sentence in this verse, in the NRSV "let it be with me according to your word" simply is semantically parallel to the first. Since Mary takes the role of a slave, she can do nothing but submit. Plummer's older commentary on Luke (1901) agrees with the classic commentary on Luke of J.M.Creed (MacMillan, London 1930, p.21), which states that "Mary humbly accepts the lot which has been appointed for her". The modern commentary on Luke by I.H.Marshall in the New International Greek Testament Commentary (Grand Rapids, MI, 1978, p.72) states that "The scene closes with Mary's humble acceptance of the will of God". The Greek clause here is almost identical to the Septuagint's Greek OT text of Genesis 31:34, and perhaps is also related to Gen 21:1 as Marshall suggests. However, the Greek text in Luke would be better translated as "let it happen to me . . . " (sic also The New Jerusalem Bible 1985, p.1687), with an aorist verb referring to a particular event rather than to a state, in which the only suggestion of any active wish is the optative mood of the verb by which Mary expresses her submission. The absence of the definite article before the word "Lord" in the Greek text is consistent with Septuagint/ NT Greek usage in which common nouns such as "kyrios" and "theos" referring to unique divinities are frequently treated as proper names.
Thus, a more accurate translation of Mary's speech in Luke 1:38 would be "Behold the slave of the Lord. Let it happen to me according to your word." Mary expresses recognition of her absolutely subordinate position and what follows from it, no less, no more.
Finally, it is worth noting that an almost identical issue about the Greek terminology for "slave" arises in the translation of the Song of Simeon, also in Luke (2:29), where the AV has "Lord, now lettest thou thy servant depart", and Goodspeed recommends and justifies a rendering as "Now, Master, you will let your slave go free, . . . " (Edgar J. GoodspeedProblems of New Testament TranslationChicago 1945, pp.77-79).