That said
« previous post | next post »
Back in June of 2002, one of William Safire's On Language columns began this way:
'The South Carolina primary between Mr. Bush and Mr. McCain in 2000," wrote Eleanor Randolph, the New York Times editorialist, referring to Representative Lindsey Graham's current campaign for the Senate, "left Republicans in his state bitter and divided. That said, both President Bush and Senator McCain have already campaigned for his election to the Senate."
In olden times, those two sentences would have been written as one, with the first clause subordinated: "Although the South Carolina primary . . . left Republicans . . . divided, both Bush and McCain . . . campaigned for his election. . . . " Or they could have remained as two sentences, with the second beginning however instead of with the voguism that said.
Immediately after introducing "[t]his absolutive participial construction now spreading like wildfire through our discourse", Safire quotes a linguist:
We turn now to Prof. John Lawler in the linguistics department of the University of Michigan: "That said is an abbreviated form of the absolutive participial phrase '(With) that (having been) said."' (You were wondering where I got that "absolutive participial" from? You think I make up this stuff? If I had taken Latin, I'd be able to explain the closely related ablative absolute.) Lawler goes on to the essential meaning of that said: "It announces a change of subject, often despite whatever was just said."
By calling that said a "voguism", and noting that things would have different in "olden times", Safire flirts with linguistic peevishness, as he often did. Many of his readers no doubt preferred to interpret him that way, just as LL commenters often write as if the question under discussion were whether a particular usage annoys them. And in this case, Safire was put onto the scent of that said by emails like "The latest abomination is the substitute for however in its many forms — having said that, that having been said, et cetera. That said, I guess I'll just have to get with it."
But after some alliterative harrumphing ("the vocabulary of vacillation"), he describes that said as an inverted-order rhetorical equivalent of "to be sure" ("X. To be sure, Y" == "Y. That said, X"), and closes with the opinion that
To be sure, that said has its good side.
Why am I bringing up this seven-year-old column? Because one of the many interesting talks that I heard at AACL 2009 was Laurel Brinton's 'The development of 'that said'".
Brinton cites the results of a COCA search showing that that said has indeed increased in popularity over the past couple of decades, apparently by moving from (some registers of) the spoken language into writing:
She also gives the results of a search of the Time Magazine archive (also available on Mark Davies excellent web site at BYU), which suggests that a fairly rapid increase in popularity beginning in the 1990s:
And she presents evidence to contradict a temporal implication of Safire's quotation of John Lawler: 'That said is an abbreviated form of the absolutive participial phrase '(With) that (having been) said."'
I suspect that John probably meant this as a description of a cognitive process, or perhaps simply as a way to clarify the grammatical structure involved; but as an account of the historical development of this phrase, Brinton argues that it is false. Before giving her own new evidence, she cites earlier authorities who give a different account:
Curme 1931 (A Grammar of the English Language. Vol. II: Syntax, 153): absolute constructions without copula are original, more modern trend is to take tense (having) and voice (having been), especially when active
Jespersen 1946 (Modern Grammar on Historical Principles. Part V: Syntax, Fourth Volume, 46, 55): that said is the original construction; the "more clumsy construction" that being said "begins to appear in the 16th century; that having been said dates from the 18th century.
Visser 1972 (An Historical Syntax of the English Language. Part Two: Syntactical Units with one Verb, 1259ff.): generally, absolute constructions with the simple past participle are original, forms with having/being/having been are introduced in the modern period and are now often preferred.
Before pursuing the history, Brinton gives a table suggesting that American and British usage may now follow somewhat different patterns:
| COCA | BNC | |
| [punc] {that, this} said, | 773 | 134 |
| [conj] {that, this} said, | 67 | 5 |
| [prep] {that, this} said, | 52 | 3 |
| all (of) {that, this} said, | 59 | 4 |
| {that, this} being said, | 172 | 16 |
| {that, this} having been said, | 36 | 4 |
| Total passive forms | 1159 | 166 |
| Frequency per million (passive forms) |
2.89 | 1.90 |
| having said {that, this} | 698 | 247 |
| Frequency per million (active forms) |
1.73 | 2.82 |
(But note that the BNC material was collected in the early 1990s, whereas COCA contains material spread equally over the period 1990-2009 — since the pattern is presumably changing on both sides of the Atlantic, an accurate account of differences will require looking across comparable sources across comparable time periods..)
She also observes that CGEL (p. 1350) analyzes that said as an "absolute construction within the class of supplements, or "elements which occupy a position in linear sequence without being integrated into the syntactic structure of the sentence", where "absolutes" express adverbial notions such as cause, condition, time, concession, manner, and attendant circumstances. She also cites the description of that said by Quirk et al. (1985) as a "contrastive-concessive disjunct".
She notes Garner's observation (in Modern American Usage, 2003) that having said that “is a frequent source of DANGLERS” when not anchored to a speaker in the main clause; he advises deleting the “casualism” as it “doesn’t say much anyway”. However, Quirk, Greenbaum & Leech's 1985 grammar (p. 623n) claims that having said that “has become so stereotyped that it can violate” the “subject-attachment rule”; and Curme (1931:158-159) also argued that in this case, the construction is no longer connected with a subject at all.
She cites the OED's gloss, added as a 2007 online supplement to the 3rd edition's entry for say ("In phrases introducing a concessive clause. having said that (also that said, that being said): even so; nevertheless") with citations only back to 1908 for that being said and 1923 for that said:
1908 Manitoba Morning Free Press (Electronic text) 1 Aug., The story of Sir James Douglas might have been told in smaller compass… That being said, James Douglas certainly deserved a place among the makers of Canada. 1923 Times 14 Aug. 5/2 The change does not appear to be popular… That said, there is little to criticize in the performance last night. 1975 A. V. GRIMSTONE in K. Sekida Zen Training 21, I believe it would be possible..to mount a convincing refutation of the argument… However, having said that, I would add that I do not believe it is really necessary to defend the practice of Zen in that way. 1986 C. SNYDER Strategic Def. Deb. 222 We have little choice; today's technology provides no alternative. That being said, we will press for radical reductions in the number and power of strategic and intermediate-range nuclear arms.
And she observes that Safire missed the chance to give the traditional name of "procatalepsis" to the rhetorical figure that he describes as follows:
Writers of opinion articles know how to use what we call a to-be-sure graph. After making an argument, some of us feel the urge to show that we are not simpletons — that we know the counterarguments, have taken them into due consideration, but still maintain our positions. In goes the to-be-sure graph, disarming our opponents with its "Yes, we know all that folderol on the other side" […]
To be sure is a rhetorical device to set up and counter the opposition after your initial point has been made. You often present the other side's position as a straw man easy to knock down and then repeat your opening argument with great force at the end. This to-be-sure trick is described by grammarians as "concessive" — that is, I'll give you this; it costs me nothing and makes me appear reasonable.
However (to use an old construction), that said is a device that works in the other direction. The point to be negated is made first: "My opponent is a great guy, a real patriot and a quick study." (End of concessive construction.) "That said, he doesn't know what he is talking about."
Compare the relevant section of Henry Peachum's 1593 Garden of Eloquence:
Procatalepsis is a forme of speech by which the Orator perceiving aforehand what might be objected against him, and hurt him, doth confute it before it be spoken, or thus: when the Orator putteth forth the same objection against himselfe, which he doth thinke his adversarie would, and then refelleth it by a reason, whereby he doth providently prevent him. Cicero: as if some Judge or commiissioner might say unto me, thou mightest have contended with a lighter action, thou mightest have come to thy right by a more easie and profitable way: wherefore either change thine action, or resist me not as Judge: or if he do prescribe after what sort I ought to sue for my right, to which objection he maketh this answere. Notwithstanding he seemeth either more fearfull then is reason a Judge should be: or else he dareth not judge that which is committed to him. Likewise against Verres, Cicero saith, that he knoweth some men will marvell, seeing so many yeares he defended many, and hurt none, he doth now come to accuse Verres, then he doth shew them that this accusation against Verres is a defence of their fellowes.
Returning to the history of that said, Brinton notes that despite the antiquity of procatalepsis, this particular way of expressing it seems to be of fairly recent origin. The earliest examples in the OED are from 1923 (that said) and 1908 (that being said), with a 1975 citation for having said that, and no examples of the other forms.
She found no examples in the following corpora of Early and Late Modern English:
Helsinki Corpus of English Texts, Modern English Section (1500–1710) 551,000 words
Corpus of English Dialogues (1560–1760) 1.2 million words
Lampeter Corpus of Early Modern English Tracts (1640–1710) 1.1 million words
Corpus of Late 18C Prose (1761–90) 300,000 words
She checked these additional corpora:
CLEMT Corpus of Late Modern English Texts (1710–1920) 10 million words
CEN Corpus of English Novels (1881–1922) 25 million words
UofV University of Virginia Electronic Text Center, Modern English Collection (1500– present)
EEPF Early English Prose Fiction (1500-1700) 200 works
18thCF Eighteenth Century Fiction (1700-1780) 96 works
Early Canadiana Online (18th c. – 1920) 3 million pages
ED English Drama (1280-1915) 3,900 plays
TIME Corpus (1923 – 2000)
She was able to find a number of earlier uses, e.g. "And these wordes sayd, she streyght her on length and rested a whyle", from 1387-88. But these are the literal usage that she calls "temporal/sequential", not the "contrastive/concessive" use. The earliest clearly concessive examples she was able to find are from the late 19th to early 20th century:
This being said, Mr. Hamilton and his colleagues of the executive council of the Territories will pardon me if I do not receive … the assurance given by them … (1894 Sessional Papers of the Dominion of Canada, Vol. 17, p. 57; ECO)
That being said we come to the question as to whether the work is worth the money (1900 Official Report of the House of Commons of the Dominion of Canada: Fifth Session, Eighth Parliament, p. 9663; ECO)
Thus contra Lawler, Brinton argues that "there is no evidence of the construction being a reduced form of (with) this/that having been said" — she found no non-recent examples of this pattern at all. The time-sequence uses of this said date from the late 16th c., with the earliest examples of temporal/sequential that said from a bit later:
This said, impatience chokes her pleading tongue (1592-93 Shakespeare, Venus and Adonis; UofV)
She is a hansome picture, And that said, all is spoken (1636 Massinger, The Great Duke of Florence; ED)
She found a few possible (but ambiguous) early examples of concessive that/this said:
They all are made my Lord, and some giue out, That ‘tis a blow giuen to religion, To weaken it, in ruining of him, That said, he neuer wisht more glorious title, Then to be call’d the scrouge of Hugenots (1608 Chapman, Charles Duke of Byron; ED)
Then, daughter, graunt me one request, To shew thou louest me as thy sisters doe, Accept a husband, whom my selfe will woo. This sayd, she cannot well deny my sute (1605 Anon., King Leir; ED)
Then there's a gap of three centuries or so, before clear examples of that said begin to appear:
The change does not appear to be popular … That said, there is little to criticize in the performance last night (1923 Times 14 Aug 5/2; OED)
So summarizing her findings in a table:
|
this said
|
that said
|
this being said
|
that being said
|
having said this
|
having said that
|
|
| Sequential | late 16th c | mid 17th c | early 16th c | — | early 17th c | mid 18th c |
| Concessive | ?17th c | 17th c?/
20th c |
late 19th c | early 20th c | mid 18th c | 20th c |
I observed in the question period after Brinton's talk that the Latin phrase his dictis ("these [things having been] said") was reasonably common — for example, it occurs six times in the Vulgate. Thus Luke 19:25
Et dixerunt ei: Domine, habet decem mnas. 26 Dico autem vobis, quia omni habenti dabitur, et abundabit: ab eo autem qui non habet, et quod habet auferetur ab eo. 27 Verumtamen inimicos meos illos, qui noluerunt me regnare super se, adducite huc: et interficite ante me. 28 Et his dictis, præcedebat ascendens Jerosolymam.
The Douay-Rheims translation:
25 And they said to him: Lord, he hath ten pounds. 26 But I say to you that to every one that hath shall be given, and he shall abound: and from him that hath not, even that which he hath shall be taken from him. 27 But as for those my enemies, who would not have me reign over them, bring them hither and kill them before me. 28 And having said these things, he went before, going up to Jerusalem.
His dictis also occurs several times in Virgil, e.g. Aeneid IV 54
His dictis impenso animum flammauit amore
spemque dedit dubiae menti soluitque pudorem.
principio delubra adeunt pacemque per aras
exquirunt;
So saying, she stirred a passion-burning breast
to love more madly still; her words infused
a doubting mind with hope, and bade the blush
of shame begone. (Theodore C. Williams translation)
Thus educated 16th- or 17th-century authors would probably be familiar with his dictis — but all the classical examples that I've been able to find are strictly temporal/sequential.
Brinton's conclusions:
The concessive meaning develops from, and has replaced, the temporal/sequential meaning.
The longer forms develop from the shorter forms.
That/this said and having said that/this have separate histories.
For the most part, the that forms have replaced the this forms.
That said shows a marked increase in frequency beginning in about 1990.
Semantic change in this construction conforms to well known paths of change.
With respect to the last point, she cites Traugott and Dasher’s “correlated paths of directionality” (Regularity in Semantic Change, 2002):
| truth-conditional | > | non-truth-conditional | ||
| content | > | content/procedural | > | procedural |
| scope within proposition | > | scope over proposition | > | scope over discourse |
| non-subjective | > | subjective | > | intersubjective |
But still, it surprises me that this process didn't take place much earlier. For example, there are plenty of Latin phrases that followed the path to "non-truth-conditional", "procedural", "scope over discourse", "intersubjective" uses. So why didn't his dictis make this journey?
And was the 1990s that said vogue a random event? Or has there been something especially procataleptic about the past couple of decades?
Martin said,
October 13, 2009 @ 7:54 am
It may be a complete red herring, but I've always wondered whether the French construction 'cela dit' is at all relevant to this…
axl said,
October 13, 2009 @ 8:47 am
I find this very interesting, but is it any more significant than trivia? For some outside profession I think it would be helpful to understand the intellectual value of this kind of detailed investigation. I guess what I mean is that I don't yet see what such an analysis might contribute to knowledge beyond the interesting facts it presents. It's an honest question – I'd like to know. And I assume there must be some high-level defence or justification of this subfield out there somewhere, right?
[(myl) Well, Laurel Brinton is in an English department, so one set of justifications would have to do with English lexicography and the history of the language. And there's a general public interest in such things, as shown by the popularity of writers like Safire.
More abstractly, there's the general question of how language and culture change. With the development of large dated archives of text and speech, it becomes possible to learn quite a bit about the distribution of such forms in space and time. In this case, it appears that the generalization of "that said" from temporal to concessive took place about 1900; but the form remained fairly rare until about 1990, when it began to increase rapidly in popularity. This apparently happened first in the spoken language — though I would guess only in fairly formal registers, and probably in certain subcultures (though the details of this are not yet clear). So this can be added to a growing collection of examples about which we know, in increasing detail, how usage changed. And this case is common enough (around 7 per million in transcripts of broadcast talk over the past decade) to make it easy to learn even more.
Finally, there's quite a bit of recent interest in applying the mathematical techniques of population genetics to language change — see e.g. Erez Lieberman, Jean-Baptiste Michel, Joe Jackson, Tina Tang & Martin A. Nowak, "Quantifying the evolutionary dynamics of language", Nature 10/11/2007. Progress in this area depends in large part on the availability of good historical data about the phenomena under study.
So you could describe Brinton's work as "botanizing in the history of the English language"; and Darwin's discoveries depended on his own decades as a botanizer, and on a couple of centuries of earlier botanizing by others.]
Acilius said,
October 13, 2009 @ 8:57 am
"That said" sounds old-fashioned to me. I'd assumed that it became popular just because it did have a quaint ring to it, so that it could be used sincerely to mark a formal register or ironically to acknowledge informality. But the data you cite suggest that I may have been wrong in that impression. Perhaps the phrase only sounds old-fashioned to me because I unthinkingly regarded it as a translation of a Latin construction. I suppose the key is in the "some registers" in the sentence "Brinton cites the results of a COCA search showing that that said has indeed increased in popularity over the past couple of decades, apparently by moving from (some registers of) the spoken language into writing."
rootlesscosmo said,
October 13, 2009 @ 9:16 am
I recall that a frequently used construction in Cicero's orations against Catiline was "quae cum ita sint," "these things being so," not exactly temporal or concessive but logically prior to the assertion that follows.
John Lawler said,
October 13, 2009 @ 10:54 am
As Mark suggested, I did in fact intend the comment "as a way to clarify the grammatical structure involved". As I recall, his research assistant's note indicated the usual anxious cluelessness, and so I just tried to give them some indication of how the phrase could be unpacked by adding stuff to make it unambiguous (if also uncolloquial). Certainly I never stated or implied that this was a historical development, for which I had up to now no evidence pro or con.
bianca steele said,
October 13, 2009 @ 11:01 am
I first noticed "that said" being used frequently on Usenet, probably between 2000 and 2001, something that should be fairly easy to research.
bianca steele said,
October 13, 2009 @ 11:31 am
"Should" being the operative word in the previous comment. Unfortunately, the Google Groups search index contains a lot of stuff that isn't actually in the database, apparently. That's even beyond needing to distinguish from "a text that said," and so on.
But, interestingly, a search of "that said" in humanities.* for 2000 results in 0 matches. The same search for 2001 results in just over 400 matches. (A combined search of both years results in over 1000 matches. I'll guess this includes duplicate posts in threads that overlapped both years.) And it drops off again after that.
Benjamin Zimmer said,
October 13, 2009 @ 12:04 pm
@bianca: The search interface for Google Groups is extremely unreliable for research of this kind. Searching on the humanities.* hierarchy for the word "said" yields a mere 14 results in 2000, 10,400 in 2001, and 109 in 2002, so something is definitely fishy here.
bianca steele said,
October 13, 2009 @ 12:58 pm
Benjamin Zimmer:
Presumably Google prioritizes allowing people to remove their posts from the Groups database over getting match counts correct (after all, they–actually Deja News, whom they acquired–didn't consult with those who'd posted to Usenet over the decades before copying their posts into the Google searchable database). Occasionally, when you request to have a post deleted for some reason, you will have to make a second request in order to remove a cached version from the index. It wouldn't be surprising if removing things from the index has surprising consequences for search counts, especially if the search/count algorithm is especially clever.
Ray Girvan said,
October 13, 2009 @ 1:33 pm
Oh. I thought this was going to be about dialect use – East Anglia, I think, of "that" as a pronoun, as in Tom Tit Tot (the English equivalent of Rumpelstiltskin):
Bill Walderman said,
October 13, 2009 @ 4:35 pm
French has a very similar absolutive construction with a concessive force: "cela dit" or "ceci dit". It might be interesting to see whether the expansion of "that said" has been paralleled by an expansion of the French equivalent.
There's apparently some disagreement over whether ceci dit is correct: http://www.etudes-litteraires.com/forum/sujet-740-ceci-dit-cela-dit
Bob Ladd said,
October 13, 2009 @ 5:48 pm
I remember noticing having said that in a concessive sense when I got to the UK in 1985, and it still seems to me more common here than that said – and I still think of it as British. Also, I'm quite happy to accept that said as an absolutive construction, whereas (pace the usage manuals you cite) having said that always triggers a momentary dangler reaction (of the sort that Geoff Pullum wrote about the other day) when I hear it.
Ian B said,
October 13, 2009 @ 10:46 pm
Re: "his dictis", my amateur guess as to the cause of its persistence as a temporal/sequential phrase is that the Latin phrase was understood to be a translation of the Greek phrase ταῦτα εἰπών, "having said these things", or something similar. This is the phrase that the Vulgate is translating at Luke 19.25, and Vergil's wording is often a translation of Homer.
Greek had both active and passive aorist/past participles, while Latin had only a passive form for the perfect/past participle. So Greek used εἰπών, a nominative aorist active participle, to sequentially connect one clause to the next and keep the same subject: e.g., "Christ said some things, and having said [εἰπών] them, he made his way up to Jerusalem." Latin, lacking an active form, was forced to translate the phrase with a perfect passive participle in an absolutive clause; but the subject of the first and second clauses were the same: e.g., "Christ said some things, and with these things having been said [dictis], he made his way up to Jerusalem."
Where the English use "that said" is absolutive and can be followed by a subject different from that of the preceding clause, the Latin use is absolutive only because it has to be to translate the Greek temporal/sequential phrase. So as a translation of a phrase that was necessarily temporal/sequential, "his dictis" might have had a harder time making the move from "contentful" to "procedural".
marie-lucie said,
October 14, 2009 @ 9:21 am
There's apparently some disagreement over whether ceci dit is correct: http://www.etudes-litteraires.com/forum/sujet-740-ceci-dit-cela-dit
As an educated French speaker I was surprised to read that ceci dit (which my mother, a teacher, was very fond of using) to refer to what one has just said might not be correct. As some commenters wrote, the distinction between (i)ci and là is so blurred nowadays that insisting that ceci dit is incorrect in the present context is pedantic.
I was going to post a reply on the site, pointing out that in conversation one would hardly say cela dit (to me a written form only) because in spoken French cela becomes ça, and no one would use ça dit in this context, but my comment was refused, apparently because my reply "3" to the test question "5 minus 2" (used where here one would get a jumble of letters to identify) was incorrect.
Karen said,
October 14, 2009 @ 11:17 am
My apartment manager is fond of sending out memos which contain an expository paragraph (e.g., we are upgrading our entry system) followed by some instructions (e.g., you must fill out this form and bring it to the office).
She invariably begins the instructions by saying "Having that been said". Invariably.
Bill Walderman said,
October 14, 2009 @ 12:02 pm
"Latin, lacking an active form, was forced to translate the phrase with a perfect passive participle in an absolutive clause;"
"Haec locutus" would translate "tauta eipo:n" or "ho:s eipo:n" into Latin. "Loquor" is a deponent verb and its past participle has an active sense. But "haec locutus" wouldn't fit into a hexameter.
'there are plenty of Latin phrases that followed the path to "non-truth-conditional", "procedural", "scope over discourse", "intersubjective" uses. So why didn't his dictis make this journey?'
Why should the fact that other Latin phrases made the journey make it necessary to explain explain the fact that his dictis didn't?
[(myl) The central point is that this development seems to be a natural one, which has taken place in a number of languages other than English; add to that the fact that Roman orators were especially fond of procatalepsis; and then observe that Latin clearly had no general linguistic or cultural prejudice against a figurative re-interpretation of phrases to turn them into rhetorical connectives. ]
Bloix said,
October 14, 2009 @ 1:20 pm
In the passage that Safire had under consideration, perhaps the intended meaning of "that said" was "nonetheless." "Nonetheless" would imply that in spite of their prior rivalry, Bush had allied themselves in support of Graham's candidacy.
But I don't know what is intended by "that said." What was previously said was that the rivalry had left the state's Republicans divided. How does that relate to the fact that Bush and McCain both campaigned for Graham? Is he saying that because of the rivalry, one would have expected Bush and McCain to disagree about Graham? What sense does that make?
To my mind, "that said" is a lazy usage. It's a rhetorical connector that's useful when the writer can't be bothered to do the thinking necessary to clarify any particular logical connection.
This kind of writing is extremely common in journalism. Grammatically correct transitions paper over the absence of logical connection all the time.
Bill Walderman said,
October 14, 2009 @ 2:37 pm
'To my mind, "that said" is a lazy usage. It's a rhetorical connector that's useful when the writer can't be bothered to do the thinking necessary to clarify any particular logical connection.'
Why is "that said" any worse than connectors like "by the same token" or "on the other hand" or "however" or just plain "but"? "That said" does prevent the writer or speaker from following up with an explanation of the logical connection of ideas, does it?
Bill Walderman said,
October 14, 2009 @ 3:11 pm
"this development seems to be a natural one, which has taken place in a number of languages other than English"
But apparently English didn't get around to it for hundreds of years–not until about 1900.
"Roman orators were especially fond of procatalepsis"
Statistics, Mr. Corpus Linguistics? (Sorry, couldn't resist.)
Italian apparently reverses the order of the words: detto questo. Here's a discussion that confirms it's used in a concessive sense in Italian:
http://prismadeformante.blogspot.com/2006/12/detto-questo.html
Kenny V said,
October 14, 2009 @ 7:29 pm
The reason that "his dictis" or "his verbis" ("with these things having been said", "with these words") and the Greek equivalents are used so much is that they signal the end of dialog and the commence of narration. Without them, the listener (or occasionally, reader) would have had a much harder time figuring out when direct speech ended. It's very useful. It is never used as "that said" is used in English, moving from one idea to another, with both expressed by the same person.
My Liberal Arts Honors college students have a strong tendency to use "that being said" rather than the shorter "that said." I have been trying to figure out why this is, but I have not met with much success.
Kenny V said,
October 14, 2009 @ 7:32 pm
as noted above, quae cum ita sint ("which things being as they are") is a much more semantically similar phrase.
Bill Walderman said,
October 15, 2009 @ 11:35 am
'The reason that "his dictis" or "his verbis" ("with these things having been said", "with these words") and the Greek equivalents are used so much is that they signal the end of dialog and the commence of narration. Without them, the listener (or occasionally, reader) would have had a much harder time figuring out when direct speech ended.'
Especially true for readers if you don't use punctuation and don't divide your words.
That said | Grumpmeister said,
January 14, 2012 @ 8:28 pm
[…] treatment of the subject (I cannot imagine why you would be), take a look at this entry on the Language Log. Advertisement GA_googleAddAttr("AdOpt", "1"); GA_googleAddAttr("Origin", "other"); […]
Concession words « English 109 said,
August 31, 2012 @ 1:58 pm
[…] obviously of course [it is true that] on the one hand…on the other hand otherwise still that said to be sure true, … but yet GA_googleAddAttr("AdOpt", "1"); GA_googleAddAttr("Origin", […]