{"id":27545,"date":"2016-08-18T06:29:10","date_gmt":"2016-08-18T11:29:10","guid":{"rendered":"http:\/\/languagelog.ldc.upenn.edu\/nll\/?p=27545"},"modified":"2016-08-18T08:14:45","modified_gmt":"2016-08-18T13:14:45","slug":"the-narrow-end-of-the-funnel","status":"publish","type":"post","link":"https:\/\/languagelog.ldc.upenn.edu\/nll\/?p=27545","title":{"rendered":"The narrow end of the funnel"},"content":{"rendered":"

<\/a>The big political story of the past 24 hours: Stephen K. Bannon<\/a>, formerly the Executive Chairman of Breitbart News<\/a>, has taken over as \"chief executive\"\u00a0of Donald Trump's presidential campaign.<\/p>\n

The big linguistic story of the past 24 hours, at least here at Language Log: an exchange between Mark Liberman and Geoff Pullum<\/a> about the rhetorical style of spontaneous speech<\/a>, as it applies to the linguistic analysis of Donald Trump's rallies.<\/p>\n

<\/p>\n

I want to continue that discussion by breaking down a few of the issues in more detail, focusing particularly on some of the ways that speech differs from writing.\u00a0As Dwight Bolinger wrote in \"Maneuvering for Stress and Intonation<\/a>\", College Composition and Communication<\/em> 1957:<\/p>\n

The first lesson that every apprentice writer must learn is what he can and what he cannot utilize out of the store of spoken devices that he has been accumulating since he learned to talk. Everyone knows that language comes out the narrow end of the funnel when it passes from speech to writing. Something is gained, no doubt, in pictographic tricks and in the precision that is made possible by our freedom to revise what we have said before anyone sets eye or ear upon it. But more is lost. All the expressiveness that we associate with a living speaker is wrung out: gesture, the look on a face, a quality of voice, the warmth of physical presence. […]<\/span><\/p>\n

The loss affects most of <\/span>all the writing that we term expository: <\/span>the essay, the historical account, the scientific treatise, where the personality <\/span>of the speaker counts for least and the logical message is everything. In drama <\/span>and novel the reader is challenged to guess at circumstances and emphases; he knows he will have to fill in. Exposi<\/span>tory writing lulls his wariness; safely ignoring the writer's tone of voice, he <\/span>feels privileged to ignore the writer's voice altogether. To lead him around the traps he thus lays for himself, the <\/span>writer must look to those lost grammati<\/span>cal ingredients. <\/span><\/span><\/p>\n

These days,\u00a0we'd prefer to use \"they\" rather than \"he\" to refer to the generic apprentice writer.<\/p>\n

And I think that Bolinger underrates \"the precision that is made possible by our freedom to revise what we have said before anyone sets eye or ear upon it\". The linguistic ingredients lost in writing\u00a0do include grammatical things, such as the ability to\u00a0use prosody to delimit\u00a0phrases, to signal their relationships, and to mark words and phrases as old or new, focus or background. And those lost linguistic ingredients also certainly include cultural and attitudinal herbs and spices.<\/p>\n

But a key fact about human speech is how unaware we generally are of its manifold disfluencies: false starts, self-corrections, ums and uhs. As speakers, we learn to recover gracefully from these stumbles, or at least some of us do. And more important,\u00a0we learn as listeners to somehow distinguish between the aspects of talking that reflect the structure and content of a\u00a0message, and the aspects that are only\u00a0symptoms of the variably-difficult\u00a0process of composition. Indeed, we learn something, or think we do, from what speakers reveal about their compositional difficulties. And sometimes speakers deploy disfluencies\u00a0with apparent\u00a0communicative intent.<\/p>\n

Linguists and psychologists\u00a0still have a lot to learn about these too-little-studied aspects of speech communication. But whatever we're doing as listeners to disentangle the threads of spoken discourse, many of those\u00a0abilities don't survive the transition to text. (Hence the complaints from reporters and transcriptionists about Donald Trump's un-telepromptered speaking style –and the fact that skilled speakers have more to learn about writing than just how to spell words and choose punctuation.)<\/p>\n

However, it should be obvious that not all styles of speaking are the same, just as not all styles of writing are the same. And this morning I want to underline one interesting thing about the style of Donald Trump's extemporized\u00a0rally speeches, namely their almost total lack of filled pauses. I'll make this point by contrasting some remarks — on rather Trumpish topics — by Stephen Bannon, from\u00a0a panel about The Future of Conservatism<\/a>\u00a0held at the National Press Club in Washington DC on 9\/24\/2013.<\/p>\n

The moderator, Nick Sorrentino<\/a>, asks:<\/p>\n

\n

what I want to do is I want to start with you Steve and
\nif you guys just wanna=
\nyou know we're talking about emerging trends in conservatism= conservatism
\num if you could just spend a couple minutes I'd like to go down the= the row here and just
\nget your thoughts initially<\/p>\n<\/div>\n

And here's Bannon's response to this first question:<\/p>\n

\n

um at= uh at Breitbart uh part of our explosive growth is because we are very focused on a= uh
\nI don't want to say anti-establishment but we believe that
\nuh when you talk about brands or you talk about the parties
\nwe don't really
\nbelieve there is a
\nfunctional conservative party in this country, we certainly don't think the Republican party is that
\nwe uh tend to look at
\nthis imperial city of Washington this boom town
\nas um they have two
\ngroups or two parties that represent the insiders'
\ncommercial party
\nand that is a collection of insider deals
\ninsider transactions uh and a= and a budding aristocracy
\nthat has made this the=
\nthe wealthiest
\nuh
\ncity
\nin the country
\num
\nand our focus is on crony capitalism
\ngrass roots support
\nsupporting tea party um
\nuh folks
\nand I think you've seen this in the current uh
\npolitical environment whether it is the
\nfood stamps bill
\nuh defeat
\nearlier in the summer
\nuh Ted Cruz's uh
\ngallant fight with Mike Lee against ObamaCare
\nuh the entire uh crony capitalism aspect of the amnesty bill
\non and on and on I think these fights
\nare gonna be the new norm
\nI mean it's gonna be a= it's really an insurgent
\nand a pop= center right populist
\nuh movement that is
\nuh virulently anti establishment
\nuh and it's going to continue to hammer
\nuh the city both the progressive left
\nand the= uh and the institutional Republican party
\nuh day in and day out and I think it's galvanizing
\nuh which is really a majority of the people the working people and middle class in this country
\nuh to really have a voice and so we see that as a huge trend and I just know from our
\nexponential growth
\num
\nthat it is something that
\nuh people are thirsting for out there, so it's a= uh we can get into more in the details later but
\neverything that we see and every trend that we see is= is very strong to a=
\nreally an outsiders'
\nuh voice and an outsiders' movement to really take their country back<\/p>\n<\/div>\n

In this passage, Bannon has 25 UHs and 5 UMs in 367 words, so that filled pauses are\u00a0100*30\/367 = 8.2% of his total word count (including filled pauses and false starts as words). In fact, UH is his commonest wordform:<\/p>\n

25 uh
\n18 and
\n18 the
\n11 that
\n10 is<\/p>\n

Now let's compare the transcribed selections of a couple of Donald Trump's speeches previously presented in LLOG posts (from \"Trump's rhetorical style<\/a>\", 9\/26\/2015, and \"The em-dash candidate<\/a>\", 8\/16\/2016).<\/p>\n

Those selections have a total of 1604 wordforms. The number of UHs and UMs? Zero.<\/p>\n

The top five\u00a0items in those passages:<\/p>\n

77 i
\n55 's
\n49 to
\n40 and
\n39 a<\/p>\n

Bannon's rate of 8.2% filled pauses is on the high side — but Trump's rate in his extemporized rally speeches is extraordinarily low. He does occasionally use filled pauses in his rallies — as I noted in\u00a0\"You know, I mean<\/a>\" (8\/14\/2016),<\/p>\n

… in my transcript of\u00a0his\u00a012\/21\/215\u00a0Grand Rapids rally<\/a> he used \"uh\" just 5 times in 11023 words, for a\u00a0rate of 0.5 per thousand words. [0.05%]<\/p>\n

And I also observed that things seem to be a bit different in his interviews, where he may be putting a bit more thought into his choice of words:<\/p>\n

In contrast, in the CNBC interview, he used \"uh\" 74 times in 5329 words, for a rate of 14 per thousand words.\u00a0[1.4%]<\/p>\n

Trump's CNBC filled-pause rate\u00a0is still relatively low \u2014 compare Hillary Clinton's Vox interview, where she used \"uh\" 210 times and \"um\" 28 times in 5985 words, for a total filled-pause rate of<\/p>\n

1000*238\/5985 or about 40 per thousand words. [4.0%]<\/p>\n

So by one measure, namely frequency of filled pauses, Donald Trump is extraordinarily fluent.<\/p>\n

I've speculated that he may have trained himself — or been trained — at some point in his life to avoid using um and uh. But it's also possible that these filled-pauses statistics\u00a0are\u00a0the consequence of a relatively unfiltered connection between his thoughts and his words.<\/p>\n

\n

Still to come: How to analyze the structure of spontaneous speech? For a small taste, consider this passage from Bannon:
\n

I mean it's gonna be a= it's really an insurgent<\/span>
\nand a pop= center right populist<\/span>
\nuh movement that is<\/span>
\nuh virulently anti establishment<\/span>
\nuh and it's going to continue to hammer<\/span>
\nuh the city both the progressive left<\/span>
\nand the= uh and the institutional Republican party<\/span>
\nuh day in and day out and I think it's galvanizing<\/span>
\nuh which is really a majority of the people the working people and middle class in this country<\/span>
\nuh to really have a voice<\/span><\/p>\n

It's clear enough in spoken form. If we run together a literal transcription of the word sequence without punctuation, we get something\u00a0that's not easy to read:<\/p>\n

I mean it's gonna be a it's really an insurgent\u00a0<\/span>and a pop- center right populist\u00a0<\/span>uh movement that is\u00a0<\/span>uh virulently anti establishment\u00a0<\/span>uh and it's going to continue to hammer\u00a0<\/span>uh the city both the progressive left\u00a0<\/span>and the uh and the institutional Republican party\u00a0<\/span>uh day in and day out and I think it's galvanizing\u00a0<\/span>uh which is really a majority of the people the working people and middle class in this country\u00a0<\/span>uh to really have a voice<\/span><\/p>\n

If we remove the filled pauses and the false starts, it gets better:<\/p>\n

I mean it's really an insurgent\u00a0<\/span>and a center right populist\u00a0<\/span>movement that is\u00a0<\/span>virulently anti establishment a<\/span>nd it's going to continue to hammer\u00a0<\/span>the city both the progressive left\u00a0<\/span>and the institutional Republican party\u00a0<\/span>day in and day out and I think it's galvanizing\u00a0<\/span>which is really a majority of the people the working people and middle class in this country t<\/span>o really have a voice<\/span><\/p>\n

Add a bit of punctuation and it's almost — but not quite — writerly:<\/p>\n

I mean it's really an insurgent\u00a0<\/span>and a center right populist\u00a0<\/span>movement that is\u00a0<\/span>virulently anti establishment, a<\/span>nd it's going to continue to hammer\u00a0<\/span>the city both the progressive left\u00a0<\/span>and the institutional Republican party,\u00a0<\/span>day in and day out, and I think it's galvanizing\u00a0<\/span>which is really a majority of the people, the working people and middle class in this country, t<\/span>o really have a voice.<\/span><\/p>\n

There are a few problems left. In particular there's the headless (or 'fused') relative clause<\/a> \"which is really a majority of the people\", here used as the object of \"galvanizing\". I believe that such clauses are used more frequently in speech than in writing, especially in constructions like this one, because in speech the otherwise-ambiguous boundaries\u00a0and function\u00a0of the clause\u00a0can be sketched prosodically.<\/p>\n","protected":false},"excerpt":{"rendered":"

The big political story of the past 24 hours: Stephen K. Bannon, formerly the Executive Chairman of Breitbart News, has taken over as \"chief executive\"\u00a0of Donald Trump's presidential campaign. The big linguistic story of the past 24 hours, at least here at Language Log: an exchange between Mark Liberman and Geoff Pullum about the rhetorical […]<\/p>\n","protected":false},"author":2,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_exactmetrics_skip_tracking":false,"_exactmetrics_sitenote_active":false,"_exactmetrics_sitenote_note":"","_exactmetrics_sitenote_category":0,"jetpack_post_was_ever_published":false,"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_memberships_contains_paid_content":false,"footnotes":""},"categories":[105,111],"tags":[],"class_list":["post-27545","post","type-post","status-publish","format-standard","hentry","category-prosody","category-rhetoric"],"jetpack_featured_media_url":"","jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/languagelog.ldc.upenn.edu\/nll\/index.php?rest_route=\/wp\/v2\/posts\/27545","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/languagelog.ldc.upenn.edu\/nll\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/languagelog.ldc.upenn.edu\/nll\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/languagelog.ldc.upenn.edu\/nll\/index.php?rest_route=\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/languagelog.ldc.upenn.edu\/nll\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=27545"}],"version-history":[{"count":25,"href":"https:\/\/languagelog.ldc.upenn.edu\/nll\/index.php?rest_route=\/wp\/v2\/posts\/27545\/revisions"}],"predecessor-version":[{"id":27570,"href":"https:\/\/languagelog.ldc.upenn.edu\/nll\/index.php?rest_route=\/wp\/v2\/posts\/27545\/revisions\/27570"}],"wp:attachment":[{"href":"https:\/\/languagelog.ldc.upenn.edu\/nll\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=27545"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/languagelog.ldc.upenn.edu\/nll\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=27545"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/languagelog.ldc.upenn.edu\/nll\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=27545"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}