“May Depend on Me Sending It You”

3.1. Grammar Writing

The production of grammars of the English language written in English started with William Bullokar’s (1586) Bref Grammar for English. Before the eighteenth century, the ongoing process of standardization was concerned mainly with orthography and lexicography, reflected in the increasing publication of spelling books and hard-word dictionaries. Michael (1970:151) argues that by the end of the seventeenth century there were just twenty-one explicitly English grammars. Grammar writing in the early eighteenth century is still modest, with fewer than forty new grammars, in stark contrast to the rapid growth during the second half of the century, which witnessed the production of two hundred works and over 1600 printings. This itself was but a preamble to the overwhelming production during the nineteenth century, with approximately 860 new grammars and over 3600 printings (Yáñez-Bouza 2016: sec. 2). Attitudes shifted over time along a continuum from a more descriptive approach in the EModE period, to the heyday of prescriptivism in the late eighteenth and early nineteenth centuries, then back again toward a more descriptive, informed approach about language usage in the late nineteenth century (on grammar writing, see Michael 1970; Tieken-Boon van Ostade 2000; Yáñez-Bouza 2016). ³

The study of historical grammars can help us understand language variation and change in the EModE and LModE periods. On the one hand, grammars can serve as evidence for the history of the English language in that they record usage, or what grammarians thought usage was like. On the other hand, there is evidence that normative grammars prescribing rules of usage have led to change: sometimes grammar rules triggered change, at times they reinforced an existing changing trend, and at other times their prescriptions had only a marginal effect (Yáñez-Bouza 2016: sec. 3). “Correct,” “proper” English was often defined in the eighteenth century according to social class and location: “the general practice of men of letters and polite speakers in the Metropolis,” as William Kenrick (1784:56) put it. Thus the peculiarities of “vulgar” speakers (uneducated, lower social class) and “provincial” speakers (the regional accents of England and the national varieties of Ireland, Scotland, and Wales) were seen as deserving stronger criticism. Note that the term dialectal had a negative connotation in this period (e.g., Sundby, Bjørge & Haugland 1991:46).

Milroy and Milroy (2012:30, 6) point out that the process of standardization is “partly aimed at preventing or inhibiting language change” and that one of the main principles invoked for this purpose is the reduction or eradication of variation. One of the consequences of this principle is that “non-standard varieties can be observed to permit more variability than standard ones” (Milroy & Milroy 2012:6); for instance, while prescriptive changes from above have (allegedly) eliminated morphological variation in the past participle of strong verbs like write (have wrote/written/writ to standard have written), morphological leveling of the past tense form of verbs with a/u alternation (sing-sang-have sung) has resisted the prescriptive forces that have attempted to impose a formal distinction, and it is still attested in some nonstandard varieties of English today (for an overview, see Yáñez-Bouza 2016 and references therein). EModE grammars were mostly concerned with morphology and word classes (Dons 2004), but LModE grammars gradually paid more attention to syntax (75 percent of the 323 grammars in the Eighteenth-Century English Grammars database contain a section on syntax/government). Given that the double object construction involves case-government and syntactic word order, that its history shows a reduction of variants, and that type (3) V–O_d–O_i such as send it her and send the book my sister is typically associated with regional usage, one might expect this feature to be discussed in grammar books and to be evaluated in terms of correctness and propriety. ⁴

3.2. Data Sample

The data examined for the purposes of this paper have been drawn from a collection of 167 works: 2 from the sixteenth century, 12 from the seventeenth century, 99 from the eighteenth century, and 54 from the nineteenth century. The reason for the unbalanced distribution is availability, as is often the case in historical corpus studies. On the one hand, grammar writing before the eighteenth century was scarce, as noted in the previous subsection, and some of these early grammars were written in Latin; the sample for the pre-1700 period is therefore inevitably small due to the lack of productivity and due to having restricted the selection to grammars written in English. Second, the Eighteenth Century Collections Online provides access to a large number of eighteenth-century works, thereby facilitating the selection of a sizeable sample for this century, but no similar resource is (yet) available for the nineteenth century. For the latter period I have consulted sources using Google Books, which presents limitations due to copyright restrictions. The data sample thus aims to contain representative sources of early normative works. The list of references selected has been drawn from Michael (1970) for the seventeenth century, the Eighteenth-Century English Grammars database for the eighteenth century, and the Collection of Nineteenth-Century Grammars for the nineteenth century. All sources have been read in search of relevant passages on any of the variants of the double object construction. When the first edition was not available—or not extant—I consulted the earliest extant edition at hand. ⁵

As reported in Section 2, the general view today is that pattern (3) V–O_d–O_i is restricted to pronominal objects and to certain regional areas of England. In order to examine whether this or any other pattern of the double object construction was “peculiar” to a certain region, the sample includes provincial glossaries and provincial dictionaries (seventeen works). Also because of the importance of dialect, an effort has been made to include authors known to be born in northern England and in Scotland, as well as works printed in northern and Scottish provincial towns (twenty-five sources related to northern England, thirty-one to Scotland). In addition, there are three French-related works by authors considered to be “naturalized English speakers” (Sundby, Bjørge & Haugland 1991:15), and a further seven related to North American colonies.

It should be noted, as an introductory observation, that overall there are about as many works which discuss double objects as those which do not: eighty-one works (48 percent) versus eighty-seven works (52 percent). However, if we take a closer look at the provincial works consulted, it turns out that none includes explicit comments on double object variation. This suggests at first sight that the narrowing process of pattern (3) V–O_d–O_i send it her to regional varieties might have been a post-1900 change. With this in mind, what follows focuses the analysis on grammar works only (n = 150), and on grammars in which at least one of the syntactic variants is discussed (n = 78).

As we can see in Table 1, the double object construction is not discussed a great deal in the seventeenth century, a period in which syntax received less attention than morphology and orthography (38.5 percent). Interest increases during the eighteenth century (50 percent), especially during the second half, when most grammars were produced (39/45 works). The construction is discussed relatively more frequently in nineteenth-century grammars, with more detailed passages and more illustrative examples (60 percent); this may well reflect the general trend of the nineteenth-century grammatical tradition: more attention to syntax, more attention to the verb, better philological background.

OPEN IN VIEWER

Table 1.

Grammar Works Examined (Raw Numbers and Percentages)

	With comments n (%)	With no comments n (%)	Total
pre-C18	5 (38.5)	8 (61.5)	13
C18	45 (50.0)	45 (50.0)	90
C19	28 (59.6)	19 (40.4)	47
Total	78 (52.0)	72 (48.0)	150

4.1. Awareness: Syntax and History

This section addresses questions (a), (b), and (c) above. The first observation to make with regard to (a) is that early grammars do comment on the four variants illustrated in types (1)-(4). The discussion of the double object construction in the works examined usually appears in sections dealing with “syntax/construction/government,” and in relation to two main topics—the omission of the preposition and verb government/case. However, the construction is treated differently.

The pattern that interests us most, type (3) V–O_d–O_i, is explicitly mentioned in just sixteen works, generally in relation to government/case (fourteen works) and only occasionally with regard to the omission of prepositions (two works). To these we can add six works in which we find illustrative examples of this pattern but without explicit remarks on its word order—one in relation to government/case and five in relation to preposition omission.

Amongst the works in which the variant (3) send it her or send the book my sister is not addressed directly, the most common topic is the omission of the preposition to/for (forty works). Word order is here hardly ever addressed explicitly; instead, authors comment on variation between prepositional and non-prepositional objects, stating that the preposition can be expressed or omitted (thirty-eight/forty works). A second topic is verb government/case (twenty-five works), usually accusative (O_d) and dative (O_i). The grammarians’ aim here is to point out that some verbs can govern two objects, without paying particular attention to the order in which they appear (fifteen works), although some authors do elaborate further on the position of the objects (ten works), most often in a descriptive way. ⁶

From a diachronic perspective there seems to be little change in terms of which topics are most often addressed. Figure 1 plots the distribution of works across three time periods and distinguishes texts in which type (3) is explicitly mentioned from those in which it is discussed only indirectly, both in relation to government/case (n = 40) and to the omission of the preposition (n = 47). The noticeable difference is that the nineteenth-century works pay relatively more attention to the double object construction as a verb-related syntactic construction (n = 21, 62 percent), whereas in the eighteenth century reference to double objects is more frequently associated with the omission of the preposition (n = 31, 66 percent).

OPEN IN VIEWER

Figure 1.

Double Objects in Early Grammars: Diachronic Trends (Raw Numbers and Percentages; N = 87)

Early grammarians are aware of syntactic variation in their own times, and a small number of them comment on the historical origin of the double object construction and on change over time (twelve works, most in the nineteenth century). The discussion often focuses on whether the O_i is a result of the omission of the preposition (e.g., send [to] me the book), or whether it is a dative case in origin. For instance, John Mulligan (1852:223) argues that nouns and pronouns “have the force of [Anglo-Saxon] datives, without any reference to the suppression of the prepositions to or for.” Allen (1824:91) also refers to the Saxon origin of the O_i, while John C. Nesfield (1898:47) and Alfred S. West (1902 [1893]:97) point to an original dative case more vaguely: me in “Give me the book” is for West “[f]ormerly a dative case with distinct inflexions [which] was used in English to express Indirect Objects.” For Edward Thring (1868:33) the origin is also a dative case but he refers to the Latin dative, not to Old English.

Alexander Crombie (1830 [1802]:270-271) shows awareness of change within the LModE period and awareness of regional variation (see research questions (b) and (c) above). He compares usage in the early eighteenth century with the early nineteenth century (see example 5). The former was a time when the two non-prepositional patterns V–O_d–O_i (3) and V–O_i–O_d (1) were used “indiscriminately,” an observation borne out by the grammarian Michael Maittaire (1712:183): “When both cases are Pronouns, ’tis no great matter, which is first.” But in Crombie’s time, variation in England has been “now reduced under one general rule,” that is pattern (3) send it her. Interestingly, the reduction in variation results in the word order which in PDE is considered regional usage, rather than in the sequence that is considered canonical nowadays, i.e., send her it (type 1). Thus, if we trust this account, there have been two diachronic changes since Crombie’s times: limited variability to pattern (3) send it her—which he reports—and another (later?) change which favors either type (1) send her it or type (2) send it to her, the PDE standard patterns. Notably, Crombie makes a further observation: the two non-prepositional patterns (1) and (3) are “still” possible in Scotland in the early nineteenth century.

4.2. Attitudes

This subsection focuses on to what extent grammars contain evaluative comments (research question (d)), and also touches on the topic of regional variation (research question (c)).

The first observation to make here is that the treatment of the double object construction in early grammars seems to be generally descriptive, in that authors notice a particular word-order pattern and provide illustrative examples, without making any further evaluation. A second general observation is that grammar works in the eighteenth and nineteenth centuries often include comments which combine grammatical correctness and propriety according to rhetorical canons. However, it seems that this particular construction of double objects is not strongly associated with the canons of rhetoric as other syntactic features were (e.g., preposition stranding and phrasal verbs), with only eleven grammars making reference to topics such as emphasis, elegance, energy, unity, or style. In fact, the analysis of attitudes reveals that there are very few remarks on correctness. In the paragraphs that follow I discuss the main strands identified in the data examined regarding descriptive, prescriptive, and proscriptive attitudes toward double objects, with a special focus on pattern (3) V–O_d–O_i send it her.

To begin with the descriptive approach, it is notable that some grammars illustrate one variant alone. In the mid-seventeenth century, Joshua Poole (1652 [1646]:22) notes that when we have two objects “the first is commonly the Dative, the latter the Accusative case,” the pattern illustrated being type (1) with pronominal O_i: “give me my booke, he promised me much, but paid me nothing.” An example from the early eighteenth century is William Turner (1710:40), where the order illustrated is also with combined pronominal O_i and nominal O_d, as in “I ask thee a question.” The same is true of the nineteenth-century grammars by John Atkin (1845:104), Ernest Adams (1871 [1868]:80; see example 6), and Charles P. Mason (1879:182-183).

Most commonly we find grammars describing variation between the present-day canonical pattern (1) V–O_i–O_d and the prepositional pattern (2) V–O_d–O_p. Authors such as William Bullokar in the late sixteenth century and Anselm Bayly in the late eighteenth century provide illustrative examples with the two objects, whether nominal or pronominal, while discussing government/case but without commenting on syntactic variation in terms of word order. Thus Bullokar’s (1586:4-5) early grammar shows combinations of type (1) “he brak me a bow, spooiled William a cot,” and type (2) “Roberd geueth a shert too Richard, and Nicolas maketh a cot for William,” which are echoed in Bayly’s (1772:66-67) work with type (1) V–O_i-pronoun–O_d-noun and type (2) V–O_d-noun–O_p-pronoun: “he gave me money, or gave money to me.” Other authors, such as Lister Metcalfe (1771 [n.d.]:79), are more explicit about the word order of the construction: the dative “is frequently placed after the Verb” and “without the Sign to, or for,” that is, the non-prepositional order in pattern (1) V–O_i-pronoun–O_d-noun. The only alternative he offers is the prepositional variant (2), as in “give me the Damsel; i.e., give the Damsel to me” and “I sent him a Letter; i.e., I sent a Letter to him.”

The descriptive approach is likewise observed in grammars which also consider pattern (3) as a variant, from the seventeenth century to the nineteenth century. Since no evaluative judgments are explicitly made, one can infer that for these grammarians sentences such as send it me are natural usage. For instance, in the seventeenth century, Guy Miège (1688:75; see example 7) notes that pronoun it occupying the O_d stands between the verb and another pronoun, thus V–O_d-it–O_i-pronoun, as in “I give it you.” This seems to be common usage to him, because in another passage about the omission of prepositions he provides similar illustrative examples with pattern (3): “bring it me.” The other pattern illustrated in his passage is type (1) with combined objects: V–O_i-pronoun–O_d-noun, as in “fill me a Glass.”

Miège was of French origin, which may well explain his remark on the French language. In the early eighteenth century another author with a French background presents pattern (3) as natural usage, too. Michael Maittaire (1712:183; see example 8) observes that when both objects are pronominal “’tis no great matter, which is first”; that is, we can say V–O_d–O_i “thou gavest them me” (type 3) as well as V–O_i–O_d “thou gavest me them” (type 1). Maittaire also shows awareness of varying patterns when the O_d is a nominal object: in combined sequences the O_i goes first: type (1) V–O_i-pronoun–O_d-noun, as in “shew us the Father.”

A similar passage is found in Lewis Brittain’s (1788:22, 69) late eighteenth-century grammar. With verbs such as give and lend, the preposition governing the dative case is “often understood,” resulting in pattern (3) V–O_d–O_i, which he illustrates with pronominal objects: “he gave it me,” “he lent it me.” This can be read as the natural word order for him, given the absence of judgmental remarks, of which Brittain has no shortage with regard to other grammatical features.

James Elphinston (1765) provides the most comprehensive account among the works examined, explaining syntactic variation by closely attending to the form of the objects. In the chapter on syntax, he notes that the preposition governing the indirect object “may” be elided in many contexts, and in fact the preposition “must” disappear in many of them (Elphinston 1765:II.153). Thus we have word order type (1) V–O_i–O_d, but also type (3) V–O_d–O_i when the O_d is occupied by the pronoun it, which Elphinston describes as “feeble”: “Give the person the thing; give it him” (Elphinston 1765:II.153). In principle, pattern (3) can be found in “all stiles,” both with O_i nominal (send it my sister) and pronominal (send it her), to the extent that the verb say will drop the preposition “in the very familiar [style] too,” that is “Say it me” (Elphinston 1765:II.154). Furthermore, in the chapter about emphasis and style, Elphinston (1765:II.210-211; see example 9) observes that the ellipsis of the preposition in the combination of nominal objects send the book my sister (type 3) is “equally familiar and elegant,” but he also acknowledges that the inverted order send my sister the book (type 1) “proves indispensable.” The exception is, again, when the O_d is pronominal with it or plural them. In this case the O_d is “naturally subjoined” to the verb before the O_i, hence V–O_d-it/them–O_i as in send it/them my sister (3). That said, in the solemn and precise style the prepositional pattern (2) send it/them to my sister is preferred.

Also in the late eighteenth century, Matthew Raine (1771:131) seems to perceive pattern (3) as a natural variant to the prepositional pattern (2), once the preposition is dropped: his fourth example in (10) illustrates V–O_d–O_i with two pronominal objects: sold it her (3), from sold it to her (2). The alternative is type (1) with the combination V–O_i-pronoun–O_d-noun (his first and third examples) or with the combination V–O_i-noun–O_d-noun (his second example). Since no evaluative comments on it are made, one can assume that the variant (3) sold it her is as natural and/or common as the others.

In the early nineteenth century, Allen (1824:91) explains that the general rule with pronominal O_is and nominal O_ds is for the O_i to precede the accusative, as in “Send me some money” (type 1: V–O_i-pronoun–O_d-noun), but he also points to pattern (3) as an exception “when that accusative is a pronoun,” such as in “You gave it me” (type 3: V–O_d-pronoun–O_i). Allen’s (1824) view is echoed at the end of the nineteenth century in Henry Sweet’s (1903 [1898]:II.16) grammar: when both objects are pronouns “the accusatival pronoun precedes,” as in “give it me!” (type 3), but when the O_d is a noun, then “a datival noun or pronoun precedes an accusatival noun”: that is, send my sister the book and send her the book (type 1: V–O_i-noun/pronoun–O_d-noun).

Although the descriptive approach seems to be common in the discussion of double objects, prescriptive comments are also made about which variant should be used in which context. It is often the case that in prescriptive passages there are references to pattern (1) and pattern (2) only, the variants characteristic of present-day standard English, but not to pattern (3), the nonstandard variant nowadays.

Amongst those who approach the topic from the perspective of word order, John Kirkby (1746:124) prescribes type (1) in the combination V–O_i-pronoun–O_d-noun, so that “the former not[es] a Person and the latter a Thing,” and the preposition is “understood,” i.e., “He gave (to) you the Money.” He adds that when we alter the order “the Preposition must be expressed,” that is, pattern (2) V–O_d–O_p, as in “She sold her House to him.” Peter W. Fogg (1796:112) makes the same point in the discussion of the omission of prepositions and the context of nominal objects: when the preposition is omitted, the pattern is the non-prepositional (1) V–O_i–O_d (“He gave not [to] God the glory”), but when it is not dropped the order “should be retransposed” to pattern (2) V–O_d–O_p (“He gave not the glory to God”).

For Samuel Oliver (1825:173) in the nineteenth century word order is a matter of correctness. He is aware of syntactic variation but the “correcter” pattern for him is type (2), as in send the book to my sister (see example 11). Nesfield (1898:47) and West (1902:97) echo the same argument in the late nineteenth century: “the Indirect object always stands first” (type 1 send her the book), Nesfield argues, but if the order is reversed “it must be preceded by the preposition ‘for’ or ‘to’” (type 2), as in “He taught Euclid (Direct) to his sons (Indirect).”

There are authors who perceive the omission of the preposition positively and thus advocate the non-prepositional pattern (1) V–O_i–O_d; there is no criticism of other patterns, just a preference for this type of syntax. For instance, John Ash (1766 [1760?]:79) argues that the preposition is “elegantly omitted” before a pronoun, as in “I give him the Book” for “I give to him the Book”—type (1) versus type (4). For John Hornsey (1793:51) prepositions are in fact “better understood than expressed”; that is, type (1) again, as in send her the book. In his treatise of rhetoric, George Campbell (1776:II.iii.IV.399) justifies the ellipsis of the preposition on the grounds that it does not cause “any inconvenience” and it is done “with a sensible advantage in point of energy”; thus, he argues, “we say, ‘Forgive us our debts,’ and not, ‘forgive to us our debts.’”

The prescriptive approach is also taken from the perspective of where the O_i must be placed, in all cases for type (1), such as send her the book and send her it: for A. Burnet (1838:103) the O_i “is always put next to the verb”; for Mulligan (1852:217) “the dative modification always precedes the objective modification”; and for Thring (1868:33) it “must come between the verb and the ordinary objective case.” None of these three authors considers variation.

Grammarians (and rhetoricians) within the prescriptive movement in grammar writing aimed at correctness and propriety, purity, precision and elegance, to name a few “ideals.” They often laid down rules about good usage. Yet, as Robert Lowth (1762:x) put it, “besides shewing what is right, the matter may be further explained by pointing out what is wrong”; attention often focused on bad usage (see, in particular, Sundby, Bjørge & Haugland 1991). In this context we would expect eighteenth- and nineteenth-century authors to proscribe certain variants of the double object construction as improper, ungrammatical, inelegant or imprecise. However, in the c.170 works examined I could only identify three grammars expressing overt criticism, and even then only mildly, of the present-day nonstandard type (3) V–O_d–O_i. These are William Ward (1765), George Crane (1843), and Hyde Clarke (1853).

In the mid-eighteenth century Ward (1765:470; see example 12) discusses the dative case with verbs that require two objects, and states that when the O_d is nominal, the O_i is placed immediately after the verb, hence type (1), as in sent me a letter, or prepositional type (2), as in sent a letter to me. But he proscribes type (3) with sequences such as sent a letter me (V–O_d-noun–O_i-pronoun).

The same argument is given a century later by the grammarian Clarke (1853:131; see example 13). He states the rule that the pronoun governed by the preposition “must” follow the verb without the preposition expressed, as in send her the book (type 1: V–O_i-pronoun–O_d-noun), but if the pronoun governed by the preposition follows the O_d, it “must take” the preposition, as in send the book to her (type 2: V–O_d-noun–O_p-pronoun). In other words, English speakers cannot say the sequence in type (3) send the book her with the combination V–O_d-noun–O_i-pronoun. (Recall that this particular combination of nominal and pronominal objects fell out of use in the late ME period, as discussed in Section 2.)

At around the same time that Clarke was writing, Crane (1843:56-57) elaborates on the syntactic patterns available in English (see example 14). In his view, the natural order is with the person O_i preceding the O_d, as in send my sister the book (type 1). Crane (1843) does allow for the variant pattern with the O_d preceding the person object, but in such a context the latter “requires” the presence of a preposition, that is, type (2) send the book to my sister (Crane 1843:56). This latter construction is in his view “better” anyway because “the complement of the person [is] connected directly by a preposition”; the absence of the preposition in examples like I wrote her a note may be ambiguous—“to her” or “for her” (Crane 1843:56). Crane is here following the rhetorical principles of precision and clarity.

Thus the recommendation in Crane’s illustrative examples in (15):

That said, in example (16) Crane shows awareness that when the O_d is the pronoun it, the word order in type (3) V–O_d-it–O_i is familiar style, e.g., send it my sister or send it her, but he insists that this pattern “is not to be recommended.” The preposition, rather, should be present; that is, he advocates pattern (2) V–O_d-it–O_p, as in send it to her.

From these three works—Ward, Clarke, Crane—we can draw two main conclusions. First, although there is some criticism, it is certainly not as severely prescriptive or proscriptive as that which other morphosyntactic features have suffered; the Dictionary of English Normative Grammar gives ample evidence of the variety of prescriptive and proscriptive epithets and their tone in the eighteenth century (Sundby, Bjørge & Haugland 1991). Second, one of the qualities that eighteenth- and nineteenth-century authors condemned the most in both writing and speech was “vulgar” and “dialectal” usage (recall Kenrick’s definition of “correct” English in Section 3), but neither quality is alluded to by Ward, Clarke, or Crane. In fact, it was noted in Section 2 that accounts of double objects in PDE tend to highlight the regional usage in England of pattern (3) send the book my sister and send it her, but none of the early normative works examined here discusses the variants of double objects in relation to English dialects. The construction is not mentioned explicitly either in dictionaries or glossaries of provincial English as a peculiar feature to any specific region of England.

In our data sample only two sources amongst those known to be related to northern England make explicit reference to pattern (3), and neither links usage to dialect. The others either do not discuss the construction at all, or they acknowledge variation but between the variants in type (1) V–O_i–O_d send my sister the book and the prepositional variants of type (2) V–O_d–O_p send the book to my sister. This suggests that pattern (3) was not a particularly salient feature in the North in Late Modern England. One of the authors here is Ward himself, from Yorkshire, who criticizes the word order V–O_d-noun–O_i-pronoun send a letter me (Ward 1765:469-470; see example (12) above), but it is known that this combination in particular had become obsolete by the EModE period (see Section 2). The other author is John Dalton (1801:116), from Cumberland, for whom pattern (3) was still an optional variant at the start of the nineteenth century: “Give it [to] me.”,—“Give it [to] James.”

Beyond regional usage in England, it is worth recalling the subtle association made by Guy Miège between the English pattern send it her and French syntax (see example 7). A century later, Hugh Mitchell (1799:54) makes a similar remark (see example 17). What is peculiar here is not only the association between the structure in English give it me and the construction in French donnez-le-moi, but also the remark about “Scotch.” Mitchell was of Scottish origin and perceives the syntactic pattern nowadays considered nonstandard dialect V–O_d–O_i send it her (3) as natural English in England, whereas the syntactic pattern that is nowadays considered standard send her it (type 1: V–O_i–O_d) is for him a Scotticism, and thus “improper,” “bad English” (Sundby, Bjørge & Haugland 1991:51). Mitchell is not alone in this, with other Scottish authors sharing the view that the norm in late eighteenth-century and nineteenth-century English usage is pattern (3) send it her. Remarks such as this appear both in works devoted to Scotticisms and in general grammar books: William Scott (1793:29) lists Give me it and Show me it as Scotticisms, and James Beattie (1797:15), Mitchell (1799:54), and William Angus (1800:xxxix) do this and also include the correct(ed) form; that is, the sequence V–O_d–O_i. ⁷

In the mid-nineteenth century, Alexander Bain (1863:181), also of Scottish origin, echoes late eighteenth-century observations, explaining that the O_i is placed last, that is, send it her; the PDE standard form send her it is for Bain “a Scotticism” (see example 18). In a different passage where Bain lists Scotticisms he rephrases an incorrect construction with the word order of type (3) as well: “I will let you see it” is changed to “Show it you” (Bain 1863:200). Further evidence from Scotland-related works suggesting that pattern (3) send it her was natural usage in the nineteenth century comes from the grammar by John M. M’Culloch (1834:146), printed in Edinburgh, where the pronominal combination object–person is provided amongst his illustrative examples: e.g., “He gave it me” (see example 19). As no further evaluative comment is added, one could again infer that for this author it is common usage to the same extent as pattern (1) illustrated in “He promised us many benefits.”

To sum up, the analysis of attitudes toward double objects reveals five main strands in the discussions documented in early writing on grammar. First, there are works which describe usage in relation to pattern (1) V–O_i–O_d and pattern (2) V–O_d–O_p alone, without making reference to pattern (3) V–O_d–O_i. Second, a number of works take pattern (3) into account alongside pattern (1) and/or pattern (2), without stating a preference for a particular variant; in other words, pattern (3), which is considered nonstandard in present-day varieties of England, was considered natural usage from the sixteenth to the nineteenth centuries. A third strand is a group of grammars which adopt a prescriptive approach by showing a preference for variant (1) or (2), again overlooking pattern (3). The fourth group of grammars is small, with just three works in which pattern (3) is overtly disfavored, if only mildly so. Finally, we have identified a group of works written by Scottish authors or printed in Scotland in which pattern (3) is in fact considered “English,” whilst pattern (1), the canonical type in PDE, is perceived as a Scotticism, hence “bad” English.