# Natural language search

**URL:** <http://discourse.iapct.org/t/natural-language-search/7509>\
**Category:** CSG2017\
**Created:** [December 9, 2017, 8:59pm UTC](http://discourse.iapct.org/t/natural-language-search/7509 "2017-12-09T20:59:28Z")\
**Posts on this page:** 14\
**Page:** 1

<div class="post-metadata">

**Author:** ![rupert](http://discourse.iapct.org/letter_avatar_proxy/v4/letter/r/f14d63/32.png) [@rupert](http://discourse.iapct.org/u/rupert)\
**Post date:** [December 9, 2017, 8:59pm UTC](http://discourse.iapct.org/t/natural-language-search/7509/1 "2017-12-09T20:59:28Z")

</div>

[From Rupert Young (2017.12.09 21.00)]

```
  Anyone know to what extent natural language processing is used in

```

google searches? They certainly don’t respond in natural language,  
but just present a list. I would have thought by now that the  
technology would be more like a chatbot. Here 's a partially  
relevant article,

> **···**
>
> ## [https://www.wired.com/2016/02/ai-is-changing-the-technology-behind-google-searches/](https://urldefense.proofpoint.com/v2/url?u=https-3A__www.wired.com_2016_02_ai-2Dis-2Dchanging-2Dthe-2Dtechnology-2Dbehind-2Dgoogle-2Dsearches_&d=DwMCaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=-dJBNItYEMOLt6aj_KjGi2LMO_Q8QB-ZzxIZIF8DGyQ&m=KC-pUlII3pbbSJH4Rqk5BAgDbA35kqq1lmn4cCRtV4A&s=pF5h-LJxaGf1oOeRii7SU9fHB2zCE1P9Rj2ZqxqRzkg&e=)
> 
> **Regards,  
> Rupert**

---

<div class="post-metadata">

**Author:** ![bnhpct](http://discourse.iapct.org/user_avatar/discourse.iapct.org/bnhpct/32/34_2.png) [@bnhpct](http://discourse.iapct.org/u/bnhpct)\
**Post date:** [December 10, 2017, 6:11pm UTC](http://discourse.iapct.org/t/natural-language-search/7509/2 "2017-12-10T18:11:25Z")

</div>

[From Bruce Nevin (2017.12.10.12:27 ET)]

Rupert Young (2017.12.09 21.00)–

There are two main approaches, statistical learning and formal linguistic analysis. They are related somewhat in the way that neural networks and control theory are related.

To the extent that a neural network reorganizes itself so as to generate realistic behavior we may assume (and verify by the Test for controlled variables) that it is controlling, and therefore that the structures resulting from its reorganization process must include negative-feedback control loops. But can those structures be examined and refined directly? Is it even possible to determine what they are?

Likewise, when a statistical ‘big data’ approach to natural-language processing is successful, the associations that it registers between words and constructions evidently correlate with those established by painstaking linguistic analysis, but are not disclosed for examination.

Fernando Pereira talks about the relationship between these two approaches to language in an important paper, “Formal grammar and information theory: together again?”, first published in 2000 in the Proceedings of the Royal Society. (A revised version was published in volume 2 of a collection that I edited, published by John Benjamins Publishing in 2002.)

> [https://www.princeton.edu/~wbialek/rome/refs/pereira\_00.pdf](https://www.princeton.edu/~wbialek/rome/refs/pereira_00.pdf)

Pereira begins as follows:

> In the last forty years, research on models of spoken and written language has been split between two seemingly irreconcilable points of view: formal linguistics in the Chomsky tradition, and information theory in the Shannon tradition. Chomsky (1957)’s famous quote signals the beginning of the split:

> 

> > > (1) Colorless green ideas sleep furiously.

> > > (2) Furiously sleep ideas green colorless.

> >

> > . . . It is fair to assume that neither sentence (1) nor (2) (nor indeed any part of these sentences) has ever occurred in an English discourse. Hence, in any statistical model for grammaticalness, these sentences will be ruled out on identical grounds as equally ‘remote’ from English. Yet (1), though nonsensical, is grammatical, while (2) is not.

> 

> Before and after the split, Zellig Harris had advocated a close alliance between grammatical and information-theoretic principles in the analysis of natural language (Harris, 1951, 1991).

He is contrasting the logical symbolic rule-processing systems of Chomsky with the empirical linguistics of Chomsky’s teacher (and mine), whose linguistic methodology I have written about in connection with PCT. This methodology is grounded in tests for controlled variables (though of course they are not called that), and identifies the controlled perceptions that constitute language.

Pereira now works for Google:

[https://sites.google.com/site/fernandopereira/](https://sites.google.com/site/fernandopereira/)

How much this informs one or another of Google’s diverse products and in what ways, is not likely to be public knowledge.

> **···**
>
> On Sat, Dec 9, 2017 at 3:59 PM, Rupert Young [rupert@perceptualrobots.com](mailto:rupert@perceptualrobots.com) wrote:
> 
> > [From Rupert Young (2017.12.09 21.00)]
> 
> > ```
> > Anyone know to what extent natural language processing is used in
> > 
> > ```
> > 
> > google searches? They certainly don’t respond in natural language,  
> > but just present a list. I would have thought by now that the  
> > technology would be more like a chatbot. Here 's a partially  
> > relevant article,  
> > [https://www.wired.com/2016/02/ai-is-changing-the-technology-behind-google-searches/](https://urldefense.proofpoint.com/v2/url?u=https-3A__www.wired.com_2016_02_ai-2Dis-2Dchanging-2Dthe-2Dtechnology-2Dbehind-2Dgoogle-2Dsearches_&d=DwMCaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=-dJBNItYEMOLt6aj_KjGi2LMO_Q8QB-ZzxIZIF8DGyQ&m=KC-pUlII3pbbSJH4Rqk5BAgDbA35kqq1lmn4cCRtV4A&s=pF5h-LJxaGf1oOeRii7SU9fHB2zCE1P9Rj2ZqxqRzkg&e=)
> 
> > –  
> > **Regards,  
> > Rupert**
> 
> >

---

<div class="post-metadata">

**Author:** ![bnhpct](http://discourse.iapct.org/user_avatar/discourse.iapct.org/bnhpct/32/34_2.png) [@bnhpct](http://discourse.iapct.org/u/bnhpct)\
**Post date:** [December 10, 2017, 11:01pm UTC](http://discourse.iapct.org/t/natural-language-search/7509/3 "2017-12-10T23:01:34Z")

</div>

[From Bruce Nevin (2017.12.10.18:00 ET)]

BTW, pace Chomsky’s influential (1957) claim, “_Colorless green ideas sleep furiously_ is at least 200,000 times as probable a sentence as _Furiously sleep ideas green colorless_.”

[http://fistfulofeuros.net/pedantry/archives/000533.html](http://fistfulofeuros.net/pedantry/archives/000533.html)

> **···**
>
> On Sun, Dec 10, 2017 at 1:11 PM, Bruce Nevin [bnhpct@gmail.com](mailto:bnhpct@gmail.com) wrote:
> 
> > [From Bruce Nevin (2017.12.10.12:27 ET)]
> 
> > Rupert Young (2017.12.09 21.00)–
> 
> > 
> 
> > There are two main approaches, statistical learning and formal linguistic analysis. They are related somewhat in the way that neural networks and control theory are related.
> 
> > 
> 
> > To the extent that a neural network reorganizes itself so as to generate realistic behavior we may assume (and verify by the Test for controlled variables) that it is controlling, and therefore that the structures resulting from its reorganization process must include negative-feedback control loops. But can those structures be examined and refined directly? Is it even possible to determine what they are?
> 
> > 
> 
> > Likewise, when a statistical ‘big data’ approach to natural-language processing is successful, the associations that it registers between words and constructions evidently correlate with those established by painstaking linguistic analysis, but are not disclosed for examination.
> 
> > 
> 
> > Fernando Pereira talks about the relationship between these two approaches to language in an important paper, “Formal grammar and information theory: together again?”, first published in 2000 in the Proceedings of the Royal Society. (A revised version was published in volume 2 of a collection that I edited, published by John Benjamins Publishing in 2002.)
> 
> > > [https://www.princeton.edu/~wbialek/rome/refs/pereira\_00.pdf](https://www.princeton.edu/~wbialek/rome/refs/pereira_00.pdf)
> 
> > Pereira begins as follows:
> 
> > 
> 
> > > In the last forty years, research on models of spoken and written language has been split between two seemingly irreconcilable points of view: formal linguistics in the Chomsky tradition, and information theory in the Shannon tradition. Chomsky (1957)’s famous quote signals the beginning of the split:
> 
> > >
> 
> > > > > (1) Colorless green ideas sleep furiously.
> 
> > > > > (2) Furiously sleep ideas green colorless.
> 
> > > >
> 
> > > > . . . It is fair to assume that neither sentence (1) nor (2) (nor indeed any part of these sentences) has ever occurred in an English discourse. Hence, in any statistical model for grammaticalness, these sentences will be ruled out on identical grounds as equally ‘remote’ from English. Yet (1), though nonsensical, is grammatical, while (2) is not.
> 
> > >
> 
> > > Before and after the split, Zellig Harris had advocated a close alliance between grammatical and information-theoretic principles in the analysis of natural language (Harris, 1951, 1991).
> 
> > 
> 
> > He is contrasting the logical symbolic rule-processing systems of Chomsky with the empirical linguistics of Chomsky’s teacher (and mine), whose linguistic methodology I have written about in connection with PCT. This methodology is grounded in tests for controlled variables (though of course they are not called that), and identifies the controlled perceptions that constitute language.
> 
> > 
> 
> > Pereira now works for Google:
> 
> > [https://sites.google.com/site/fernandopereira/](https://sites.google.com/site/fernandopereira/)
> 
> > How much this informs one or another of Google’s diverse products and in what ways, is not likely to be public knowledge.
> 
> > 
> 
> > > /Bruce
> 
> > 
> 
> > 
> 
> > 
> 
> On Sat, Dec 9, 2017 at 3:59 PM, Rupert Young [rupert@perceptualrobots.com](mailto:rupert@perceptualrobots.com) wrote:
> 
> > [From Rupert Young (2017.12.09 21.00)]
> 
> > ```
> > Anyone know to what extent natural language processing is used in
> > 
> > ```
> > 
> > google searches? They certainly don’t respond in natural language,  
> > but just present a list. I would have thought by now that the  
> > technology would be more like a chatbot. Here 's a partially  
> > relevant article,  
> > [https://www.wired.com/2016/02/ai-is-changing-the-technology-behind-google-searches/](https://urldefense.proofpoint.com/v2/url?u=https-3A__www.wired.com_2016_02_ai-2Dis-2Dchanging-2Dthe-2Dtechnology-2Dbehind-2Dgoogle-2Dsearches_&d=DwMCaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=-dJBNItYEMOLt6aj_KjGi2LMO_Q8QB-ZzxIZIF8DGyQ&m=KC-pUlII3pbbSJH4Rqk5BAgDbA35kqq1lmn4cCRtV4A&s=pF5h-LJxaGf1oOeRii7SU9fHB2zCE1P9Rj2ZqxqRzkg&e=)
> 
> > –  
> > **Regards,  
> > Rupert**
> 
> >

---

<div class="post-metadata">

**Author:** ![rupert](http://discourse.iapct.org/letter_avatar_proxy/v4/letter/r/f14d63/32.png) [@rupert](http://discourse.iapct.org/u/rupert)\
**Post date:** [December 11, 2017, 2:22pm UTC](http://discourse.iapct.org/t/natural-language-search/7509/4 "2017-12-11T14:22:06Z")

</div>

[From Rupert Young (2017.12.11  
14.25)]

```
Thanks. Fyi, I didn't get the post below, though see it did appear

```

on the archive.  
Rupert

> **···**
>
> On 10/12/2017 23:01, Bruce Nevin wrote:
> 
> > [From Bruce Nevin (2017.12.10.18:00 ET)]
> 
> > BTW, pace Chomsky’s influential (1957) claim, “\* Colorless  
> > green ideas sleep furiously\* is  
> > at least 200,000 times as probable a sentence as \* Furiously  
> > sleep ideas green colorless\*.”
> 
> > [http://fistfulofeuros.net/pedantry/archives/000533.html](https://urldefense.proofpoint.com/v2/url?u=http-3A__fistfulofeuros.net_pedantry_archives_000533.html&d=DwMDaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=-dJBNItYEMOLt6aj_KjGi2LMO_Q8QB-ZzxIZIF8DGyQ&m=lAkcdJ3QGNMCYhy4e7I4-ylglUE0z6PCKfLmq6WJeOw&s=Uq1kjwgkifuUC466YvytTF27vGLZKf95JT10m9sU3lg&e=)
> 
> > ```
> > On Sun, Dec 10, 2017 at 1:11 PM, Bruce
> > 
> > ```
> > 
> > Nevin [bnhpct@gmail.com](mailto:bnhpct@gmail.com)  
> > wrote:
> > 
> > > [From Bruce Nevin (2017.12.10.12:27 ET)]
> 
> > > ```
> > > Rupert Young
> > > 
> > > ```
> > > 
> > > (2017.12.09 21.00)–
> 
> > >
> 
> > > ```
> > > There are two main
> > > 
> > > ```
> > > 
> > > approaches, statistical learning and formal linguistic  
> > > analysis. They are related somewhat in the way that  
> > > neural networks and control theory are related.
> 
> > >
> 
> > > ```
> > > To the extent that a
> > > 
> > > ```
> > > 
> > > neural network reorganizes itself so as to generate  
> > > realistic behavior we may assume (and verify by the  
> > > Test for controlled variables) that it is controlling,  
> > > and therefore that the structures resulting from its  
> > > reorganization process must include negative-feedback  
> > > control loops. But can those structures be examined  
> > > and refined directly? Is it even possible to determine  
> > > what they are?
> 
> > >
> 
> > > ```
> > > Likewise, when a
> > > 
> > > ```
> > > 
> > > statistical ‘big data’ approach to natural-language  
> > > processing is successful, the associations that it  
> > > registers between words and constructions evidently  
> > > correlate with those established by painstaking  
> > > linguistic analysis, but are not disclosed for  
> > > examination.
> 
> > >
> 
> > > Fernando Pereira talks about the relationship  
> > > between these two approaches to language in an  
> > > important paper, “Formal grammar and information  
> > > theory: together again?”, first published in 2000 in  
> > > the Proceedings of the Royal Society. (A revised  
> > > version was published in volume 2 of a collection that  
> > > I edited, published by John Benjamins Publishing in  
> > > 2002.)
> 
> > > > [https://www.princeton.edu/~wbialek/rome/refs/pereira\_00.pdf](https://urldefense.proofpoint.com/v2/url?u=https-3A__www.princeton.edu_-257Ewbialek_rome_refs_pereira-5F00.pdf&d=DwMDaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=-dJBNItYEMOLt6aj_KjGi2LMO_Q8QB-ZzxIZIF8DGyQ&m=lAkcdJ3QGNMCYhy4e7I4-ylglUE0z6PCKfLmq6WJeOw&s=NVnCBUOGCv2D53CASLPsLHix8jBfxtwxgyRflfYW8m4&e=)
> 
> > > ```
> > > Pereira begins as
> > > 
> > > ```
> > > 
> > > follows:
> 
> > >
> 
> > > > ```
> > > > In the last forty years, research on models of
> > > > 
> > > > ```
> > > > 
> > > > spoken and written language has been split between  
> > > > two seemingly irreconcilable points of view: formal  
> > > > linguistics in the Chomsky tradition, and  
> > > > information theory in the Shannon tradition. Chomsky  
> > > > (1957)’s famous quote signals the beginning of the  
> > > > split:
> 
> > > >
> 
> > > > > > (1) Colorless green ideas sleep furiously.
> 
> > > > > > (2) Furiously sleep ideas green colorless.
> 
> > > > >
> 
> > > > > ```
> > > > > . . . It is fair to assume that neither
> > > > > 
> > > > > ```
> > > > > 
> > > > > sentence (1) nor (2) (nor indeed any part of these  
> > > > > sentences) has ever occurred in an English  
> > > > > discourse. Hence, in any statistical model for  
> > > > > grammaticalness, these sentences will be ruled out  
> > > > > on identical grounds as equally ‘remote’ from  
> > > > > English. Yet (1), though nonsensical, is  
> > > > > grammatical, while (2) is not.
> 
> > > >
> 
> > > > ```
> > > > Before and after the split, Zellig Harris had
> > > > 
> > > > ```
> > > > 
> > > > advocated a close alliance between grammatical and  
> > > > information-theoretic principles in the analysis of  
> > > > natural language (Harris, 1951, 1991).
> 
> > >
> 
> > > ```
> > > He is contrasting the logical symbolic
> > > 
> > > ```
> > > 
> > > rule-processing systems of Chomsky with the empirical  
> > > linguistics of Chomsky’s teacher (and mine), whose linguistic methodology I  
> > > have written about in connection with PCT. This  
> > > methodology is grounded in tests for controlled  
> > > variables (though of course they are not called  
> > > that), and identifies the controlled  
> > > perceptions that constitute language.
> 
> > >
> 
> > > Pereira now works  
> > > for Google:
> 
> > > [https://sites.google.com/site/fernandopereira/](https://urldefense.proofpoint.com/v2/url?u=https-3A__sites.google.com_site_fernandopereira_&d=DwMDaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=-dJBNItYEMOLt6aj_KjGi2LMO_Q8QB-ZzxIZIF8DGyQ&m=lAkcdJ3QGNMCYhy4e7I4-ylglUE0z6PCKfLmq6WJeOw&s=vM2l9mWGDEWIl41GeElukDKc_Cw4tMQD3EPeDueTf5w&e=)
> 
> > > ```
> > > How much this informs one or another of Google's
> > > 
> > > ```
> > > 
> > > diverse products and in what ways, is not likely to be  
> > > public knowledge.
> 
> > >
> 
> > > > /Bruce
> 
> > >
> 
> > >
> 
> > >
> 
> > > ```
> > > On Sat, Dec 9, 2017 at 3:59
> > > 
> > > ```
> > > 
> > > PM, Rupert Young [rupert@perceptualrobots.com](mailto:rupert@perceptualrobots.com)  
> > > wrote:
> > > 
> > > > [From Rupert Young (2017.12.09 21.00)]
> 
> > > > ```
> > > > Anyone know to what extent natural language
> > > > 
> > > > ```
> > > > 
> > > > processing is used in google searches? They  
> > > > certainly don’t respond in natural language,  
> > > > but just present a list. I would have thought  
> > > > by now that the technology would be more like  
> > > > a chatbot. Here 's a partially relevant  
> > > > article,  
> > > > [https://www.wired.com/2016/02/ai-is-changing-the-technology-behind-google-searches/](https://urldefense.proofpoint.com/v2/url?u=https-3A__www.wired.com_2016_02_ai-2Dis-2Dchanging-2Dthe-2Dtechnology-2Dbehind-2Dgoogle-2Dsearches_&d=DwMCaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=-dJBNItYEMOLt6aj_KjGi2LMO_Q8QB-ZzxIZIF8DGyQ&m=KC-pUlII3pbbSJH4Rqk5BAgDbA35kqq1lmn4cCRtV4A&s=pF5h-LJxaGf1oOeRii7SU9fHB2zCE1P9Rj2ZqxqRzkg&e=)
> 
> > > > ```
> > > > --
> > > > 
> > > > ```
> > > > 
> > > > **Regards,  
> > > > Rupert**
> 
> > > >
> 
> > >
> 
> >

---

<div class="post-metadata">

**Author:** ![rupert](http://discourse.iapct.org/letter_avatar_proxy/v4/letter/r/f14d63/32.png) [@rupert](http://discourse.iapct.org/u/rupert)\
**Post date:** [December 12, 2017, 9:32am UTC](http://discourse.iapct.org/t/natural-language-search/7509/5 "2017-12-12T09:32:46Z")

</div>

[From Rupert Young (2017.12.12 09.30)]

```
  I'll read the paper. Do we have anything that puts language

```

(textual) in the PCT framework? That is, in the context of  
separating (parsing) it into perceptual types at the different  
levels in the hierarchy.

Regards,

Rupert

(Bruce Nevin (2017.12.10.12:27 ET)]

> **···**
>
> > 
> 
> > ```
> > Rupert Young
> > 
> > ```
> > 
> > (2017.12.09 21.00)–
> 
> > 
> 
> > ```
> > There are two main
> > 
> > ```
> > 
> > approaches, statistical learning and formal linguistic  
> > analysis. They are related somewhat in the way that  
> > neural networks and control theory are related.
> 
> > 
> 
> > ```
> > To the extent that a
> > 
> > ```
> > 
> > neural network reorganizes itself so as to generate  
> > realistic behavior we may assume (and verify by the  
> > Test for controlled variables) that it is controlling,  
> > and therefore that the structures resulting from its  
> > reorganization process must include negative-feedback  
> > control loops. But can those structures be examined  
> > and refined directly? Is it even possible to determine  
> > what they are?
> 
> > 
> 
> > ```
> > Likewise, when a
> > 
> > ```
> > 
> > statistical ‘big data’ approach to natural-language  
> > processing is successful, the associations that it  
> > registers between words and constructions evidently  
> > correlate with those established by painstaking  
> > linguistic analysis, but are not disclosed for  
> > examination.
> 
> > 
> 
> > Fernando Pereira talks about the relationship  
> > between these two approaches to language in an  
> > important paper, “Formal grammar and information  
> > theory: together again?”, first published in 2000 in  
> > the Proceedings of the Royal Society. (A revised  
> > version was published in volume 2 of a collection that  
> > I edited, published by John Benjamins Publishing in  
> > 2002.)
> 
> > > [https://www.princeton.edu/~wbialek/rome/refs/pereira\_00.pdf](https://urldefense.proofpoint.com/v2/url?u=https-3A__www.princeton.edu_-257Ewbialek_rome_refs_pereira-5F00.pdf&d=DwMDaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=-dJBNItYEMOLt6aj_KjGi2LMO_Q8QB-ZzxIZIF8DGyQ&m=K7zFtBXN0l9Q5Njq8LUpSS9SY728D6VuwnJvYxka518&s=9h3_hoTOCNN-ZLA_SPAGToBFtMDJVNpbnntPFr_Ci-I&e=)
> 
> > ```
> > Pereira begins as
> > 
> > ```
> > 
> > follows:
> 
> > 
> 
> > > ```
> > > In the last forty years, research on models of
> > > 
> > > ```
> > > 
> > > spoken and written language has been split between  
> > > two seemingly irreconcilable points of view: formal  
> > > linguistics in the Chomsky tradition, and  
> > > information theory in the Shannon tradition. Chomsky  
> > > (1957)’s famous quote signals the beginning of the  
> > > split:
> 
> > >
> 
> > > > > (1) Colorless green ideas sleep furiously.
> 
> > > > > (2) Furiously sleep ideas green colorless.
> 
> > > >
> 
> > > > ```
> > > > . . . It is fair to assume that neither
> > > > 
> > > > ```
> > > > 
> > > > sentence (1) nor (2) (nor indeed any part of these  
> > > > sentences) has ever occurred in an English  
> > > > discourse. Hence, in any statistical model for  
> > > > grammaticalness, these sentences will be ruled out  
> > > > on identical grounds as equally ‘remote’ from  
> > > > English. Yet (1), though nonsensical, is  
> > > > grammatical, while (2) is not.
> 
> > >
> 
> > > ```
> > > Before and after the split, Zellig Harris had
> > > 
> > > ```
> > > 
> > > advocated a close alliance between grammatical and  
> > > information-theoretic principles in the analysis of  
> > > natural language (Harris, 1951, 1991).
> 
> > 
> 
> > ```
> > He is contrasting the logical symbolic
> > 
> > ```
> > 
> > rule-processing systems of Chomsky with the empirical  
> > linguistics of Chomsky’s teacher (and mine), whose linguistic methodology I  
> > have written about in connection with PCT. This  
> > methodology is grounded in tests for controlled  
> > variables (though of course they are not called  
> > that), and identifies the controlled  
> > perceptions that constitute language.
> 
> > 
> 
> > Pereira now works  
> > for Google:
> 
> > [https://sites.google.com/site/fernandopereira/](https://urldefense.proofpoint.com/v2/url?u=https-3A__sites.google.com_site_fernandopereira_&d=DwMDaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=-dJBNItYEMOLt6aj_KjGi2LMO_Q8QB-ZzxIZIF8DGyQ&m=K7zFtBXN0l9Q5Njq8LUpSS9SY728D6VuwnJvYxka518&s=gRGcvG78wHWLkmBC3tsBixNf2LP9Px7K46kciMwb6Mg&e=)
> 
> > ```
> > How much this informs one or another of Google's
> > 
> > ```
> > 
> > diverse products and in what ways, is not likely to be  
> > public knowledge.
> 
> > 
> 
> > > /Bruce
> 
> > 
> 
> > 
> 
> > 
> 
> ```
> On Sat, Dec 9, 2017 at 3:59
> 
> ```
> 
> PM, Rupert Young [rupert@perceptualrobots.com](mailto:rupert@perceptualrobots.com)  
> wrote:
> 
> > [From Rupert Young (2017.12.09 21.00)]
> 
> > ```
> > Anyone know to what extent natural language
> > 
> > ```
> > 
> > processing is used in google searches? They  
> > certainly don’t respond in natural language,  
> > but just present a list. I would have thought  
> > by now that the technology would be more like  
> > a chatbot. Here 's a partially relevant  
> > article,  
> > [https://www.wired.com/2016/02/ai-is-changing-the-technology-behind-google-searches/](https://urldefense.proofpoint.com/v2/url?u=https-3A__www.wired.com_2016_02_ai-2Dis-2Dchanging-2Dthe-2Dtechnology-2Dbehind-2Dgoogle-2Dsearches_&d=DwMCaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=-dJBNItYEMOLt6aj_KjGi2LMO_Q8QB-ZzxIZIF8DGyQ&m=KC-pUlII3pbbSJH4Rqk5BAgDbA35kqq1lmn4cCRtV4A&s=pF5h-LJxaGf1oOeRii7SU9fHB2zCE1P9Rj2ZqxqRzkg&e=)
> 
> > ```
> > --
> > 
> > ```
> > 
> > **Regards,  
> > Rupert**
> 
> >

---

<div class="post-metadata">

**Author:** ![bnhpct](http://discourse.iapct.org/user_avatar/discourse.iapct.org/bnhpct/32/34_2.png) [@bnhpct](http://discourse.iapct.org/u/bnhpct)\
**Post date:** [December 15, 2017, 5:07pm UTC](http://discourse.iapct.org/t/natural-language-search/7509/6 "2017-12-15T17:07:17Z")

</div>

[From Bruce Nevin (2012.12.15.12:00 ET)]

Rupert Young (2017.12.12 09.30)–

> RY: Do we have anything that puts language (textual) in the PCT framework? That is, in the context of separating (parsing) it into perceptual types at the different levels in the hierarchy.

I have three things that address this. My chapter in LCS IV I am forbidden to distribute until after publication. The following derive from a conference in Duino sponsored by the University of Paris.

[http://zelligharris.org/Embodied.grammar.pdf](http://zelligharris.org/Embodied.grammar.pdf)

[http://zelligharris.org/Nevin20160521.pdf](http://zelligharris.org/Nevin20160521.pdf)

The presentation that I gave at Stanford in February 2012:

[http://www.livingcontrolsystems.com/SPLaT/SPLaT\_Nevin.html](http://www.livingcontrolsystems.com/SPLaT/SPLaT_Nevin.html)

This concerns meanings only at the most primitive level of phonemic contrast: speakers and hearers control the distinctness of different words from one another. Kinesthetic and tactile perceptions of pronunciation are controlled in real time; their acoustic effects can only be controlled over time as contrasting words and syllables are repeated, and they are controlled by ‘tuning’ the references for the real-time kinesthetic and tactile perceptions. A proposed PCT model indicates a solution to the puzzle. A working simulation has not yet been built. It would presumably involve some kind of numerical surrogate for human speech apparatus. This engages only the lowest levels of language structure and of the perceptual hierarchy. (My host introducing me was Ivan Sag, since deceased.) I just noticed that the copy that I had uploaded to youtube is gone (only a snip of discussion at the end was there), so I put it up again.

There are also two papers in the 2003 Festschrift for Bill:

[http://iapct.org/festschrift/judd.html](http://iapct.org/festschrift/judd.html)

[http://iapct.org/festschrift/nevin.html](http://iapct.org/festschrift/nevin.html)

> **···**
>
> On Tue, Dec 12, 2017 at 4:32 AM, Rupert Young [rupert@perceptualrobots.com](mailto:rupert@perceptualrobots.com) wrote:
> 
> > [From Rupert Young (2017.12.12 09.30)]
> 
> > ```
> > I'll read the paper. Do we have anything that puts language
> > 
> > ```
> > 
> > (textual) in the PCT framework? That is, in the context of  
> > separating (parsing) it into perceptual types at the different  
> > levels in the hierarchy.
> 
> > Regards,
> 
> > Rupert
> 
> > 
> 
> > ```
> > (Bruce Nevin (2017.12.10.12:27 ET)]
> > 
> > ```
> 
> > 
> 
> > ```
> > Rupert Young
> > 
> > ```
> > 
> > (2017.12.09 21.00)–
> 
> > 
> 
> > ```
> > There are two main
> > 
> > ```
> > 
> > approaches, statistical learning and formal linguistic  
> > analysis. They are related somewhat in the way that  
> > neural networks and control theory are related.
> 
> > 
> 
> > ```
> > To the extent that a
> > 
> > ```
> > 
> > neural network reorganizes itself so as to generate  
> > realistic behavior we may assume (and verify by the  
> > Test for controlled variables) that it is controlling,  
> > and therefore that the structures resulting from its  
> > reorganization process must include negative-feedback  
> > control loops. But can those structures be examined  
> > and refined directly? Is it even possible to determine  
> > what they are?
> 
> > 
> 
> > ```
> > Likewise, when a
> > 
> > ```
> > 
> > statistical ‘big data’ approach to natural-language  
> > processing is successful, the associations that it  
> > registers between words and constructions evidently  
> > correlate with those established by painstaking  
> > linguistic analysis, but are not disclosed for  
> > examination.
> 
> > 
> 
> > Fernando Pereira talks about the relationship  
> > between these two approaches to language in an  
> > important paper, “Formal grammar and information  
> > theory: together again?”, first published in 2000 in  
> > the Proceedings of the Royal Society. (A revised  
> > version was published in volume 2 of a collection that  
> > I edited, published by John Benjamins Publishing in  
> > 2002.)
> 
> > > [https://www.princeton.edu/~wbialek/rome/refs/pereira\_00.pdf](https://urldefense.proofpoint.com/v2/url?u=https-3A__www.princeton.edu_-257Ewbialek_rome_refs_pereira-5F00.pdf&d=DwMDaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=-dJBNItYEMOLt6aj_KjGi2LMO_Q8QB-ZzxIZIF8DGyQ&m=K7zFtBXN0l9Q5Njq8LUpSS9SY728D6VuwnJvYxka518&s=9h3_hoTOCNN-ZLA_SPAGToBFtMDJVNpbnntPFr_Ci-I&e=)
> 
> > ```
> > Pereira begins as
> > 
> > ```
> > 
> > follows:
> 
> > 
> 
> > > ```
> > > In the last forty years, research on models of
> > > 
> > > ```
> > > 
> > > spoken and written language has been split between  
> > > two seemingly irreconcilable points of view: formal  
> > > linguistics in the Chomsky tradition, and  
> > > information theory in the Shannon tradition. Chomsky  
> > > (1957)’s famous quote signals the beginning of the  
> > > split:
> 
> > >
> 
> > > > > (1) Colorless green ideas sleep furiously.
> 
> > > > > (2) Furiously sleep ideas green colorless.
> 
> > > >
> 
> > > > ```
> > > > . . . It is fair to assume that neither
> > > > 
> > > > ```
> > > > 
> > > > sentence (1) nor (2) (nor indeed any part of these  
> > > > sentences) has ever occurred in an English  
> > > > discourse. Hence, in any statistical model for  
> > > > grammaticalness, these sentences will be ruled out  
> > > > on identical grounds as equally ‘remote’ from  
> > > > English. Yet (1), though nonsensical, is  
> > > > grammatical, while (2) is not.
> 
> > >
> 
> > > ```
> > > Before and after the split, Zellig Harris had
> > > 
> > > ```
> > > 
> > > advocated a close alliance between grammatical and  
> > > information-theoretic principles in the analysis of  
> > > natural language (Harris, 1951, 1991).
> 
> > 
> 
> > ```
> > He is contrasting the logical symbolic
> > 
> > ```
> > 
> > rule-processing systems of Chomsky with the empirical  
> > linguistics of Chomsky’s teacher (and mine), whose linguistic methodology I  
> > have written about in connection with PCT. This  
> > methodology is grounded in tests for controlled  
> > variables (though of course they are not called  
> > that), and identifies the controlled  
> > perceptions that constitute language.
> 
> > 
> 
> > Pereira now works  
> > for Google:
> 
> > [https://sites.google.com/site/fernandopereira/](https://urldefense.proofpoint.com/v2/url?u=https-3A__sites.google.com_site_fernandopereira_&d=DwMDaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=-dJBNItYEMOLt6aj_KjGi2LMO_Q8QB-ZzxIZIF8DGyQ&m=K7zFtBXN0l9Q5Njq8LUpSS9SY728D6VuwnJvYxka518&s=gRGcvG78wHWLkmBC3tsBixNf2LP9Px7K46kciMwb6Mg&e=)
> 
> > ```
> > How much this informs one or another of Google's
> > 
> > ```
> > 
> > diverse products and in what ways, is not likely to be  
> > public knowledge.
> 
> > 
> 
> > > /Bruce
> 
> > 
> 
> > 
> 
> > 
> 
> ```
> On Sat, Dec 9, 2017 at 3:59
> 
> ```
> 
> PM, Rupert Young [rupert@perceptualrobots.com](mailto:rupert@perceptualrobots.com)  
> wrote:
> 
> > [From Rupert Young (2017.12.09 21.00)]
> 
> > ```
> > Anyone know to what extent natural language
> > 
> > ```
> > 
> > processing is used in google searches? They  
> > certainly don’t respond in natural language,  
> > but just present a list. I would have thought  
> > by now that the technology would be more like  
> > a chatbot. Here 's a partially relevant  
> > article,  
> > [https://www.wired.com/2016/02/ai-is-changing-the-technology-behind-google-searches/](https://urldefense.proofpoint.com/v2/url?u=https-3A__www.wired.com_2016_02_ai-2Dis-2Dchanging-2Dthe-2Dtechnology-2Dbehind-2Dgoogle-2Dsearches_&d=DwMCaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=-dJBNItYEMOLt6aj_KjGi2LMO_Q8QB-ZzxIZIF8DGyQ&m=KC-pUlII3pbbSJH4Rqk5BAgDbA35kqq1lmn4cCRtV4A&s=pF5h-LJxaGf1oOeRii7SU9fHB2zCE1P9Rj2ZqxqRzkg&e=)
> 
> > ```
> > --
> > 
> > ```
> > 
> > **Regards,  
> > Rupert**
> 
> >

---

<div class="post-metadata">

**Author:** ![MartinT](http://discourse.iapct.org/letter_avatar_proxy/v4/letter/m/ee7513/32.png) [@MartinT](http://discourse.iapct.org/u/MartinT)\
**Post date:** [December 15, 2017, 6:49pm UTC](http://discourse.iapct.org/t/natural-language-search/7509/7 "2017-12-15T18:49:11Z")

</div>

[Martin Taylor 2017.12.15.13.14]

> [From Rupert Young (2017.12.12 09.30)]

> ```
> ...Do we have anything that puts language (textual) in the PCT
> 
> ```
> 
> framework? …

> 

```
I do, but since it is a substantial part (roughly 100 single-spaced

```

pages out of 600 at the moment) of the book I am writing, I don’t  
think I can post it, though I might possibly be able to provide  
drafts to individuals who have the background and who would be  
prepared to offer helpful comments. All the same, I think it not  
unreasonable to post a small section that is a sort of introduction  
to the language part. Here is a PCT “just So Story” about how a baby  
produces and mother learns a first word.

```
The working title of the book is "* Powers of Perceptual Control: a

```

PCT enquiry into Language, Culture, Power, and Politics\* ", pun  
intended. My LCS IV chapter is more on the mechanism of protocols,  
and not really relevant to the question. However, some of my old  
Layered Protocol papers and chapters might be relevant, since LPT is  
just a special case of PCT invented before I knew of PCT. A few are  
on researchgate.

```
Martin

```

[12.1CoraAndIvan.pdf](http://discourse.iapct.org/uploads/short-url/6iErHZWha20wzx49VHJAuDbVua3.pdf) (177 KB)

---

<div class="post-metadata">

**Author:** ![bnhpct](http://discourse.iapct.org/user_avatar/discourse.iapct.org/bnhpct/32/34_2.png) [@bnhpct](http://discourse.iapct.org/u/bnhpct)\
**Post date:** [December 16, 2017, 11:00pm UTC](http://discourse.iapct.org/t/natural-language-search/7509/8 "2017-12-16T23:00:49Z")

</div>

[From Bruce Nevin (2017.12.16.18:00 ET)]

Martin Taylor 2017.12.15.13.14 –

It might be useful to have your definition of ‘protocol’, Martin.

In the usual definitions, a protocol is (or specifies) constraints to which a communicative interaction must conform in order to be perceived as such. Presupposed metacommunication, if you will. Here are two definitions (for diplomacy and for computer intercommunication, respectively) from the Random House Dictionary:

```
1. the customs and regulations dealing with diplomatic formality, precedence, and etiquette.

[...]

7. Computers. a set of rules governing the format of messages that are exchanged between computers.

```

Thus, the Internet Protocol specifies a number of things (such as IP addresses) to which communications from one computer across the Internet to another must comply.

You refer to the baby’s initiating protocol and the mother’s continuation protocol. The word here seems to refer to parts of an instance of communication rather than to the whole. Maybe this is analogous to the definitions 2-5 elided above. They are:

```
2. an original draft, minute, or record from which a document, esp. a treaty, is prepared.

3. a supplementary international agreement.

4. an agreement between states.

5. an annex to a treaty giving data relating to it.

```

In these cases, it’s not synecdoche–the part taken to represent the whole–but rather the word for the whole (‘protocol’) is used to refer to a part. The analogy breaks down because these are parts of a particular instance of protocol-compliant communication (negotiation of a treaty, etc.), and your terms initiating protocol and continuation protocol are at a higher level of generalization. An initiating protocol is any observed behavior by party A that is [intended to be] perceived by party B as initiating an an exchange compliant with a known protocol; and so on.

There are two other dictionary definitions which I believe are a bit too divergent to be relevant:

```
6. Med. the plan for carrying out a scientific study or a patient's treatment regimen.

8. Also called protocol statement, protocol sentence, protocol proposition. Philos. a statement reporting an observation or experience in the most fundamental terms without interpretation: sometimes taken as the basis of empirical verification, as of scientific laws.

```

> **···**
>
> On Fri, Dec 15, 2017 at 1:49 PM, Martin Taylor [mmt-csg@mmtaylor.net](mailto:mmt-csg@mmtaylor.net) wrote:
> 
> > [Martin Taylor 2017.12.15.13.14]
> 
> > > [From Rupert Young (2017.12.12 09.30)]
> 
> > > ```
> > > ...Do we have anything that puts language (textual) in the PCT
> > > 
> > > ```
> > > 
> > > framework? …
> 
> > > 
> > 
> > ```
> > I do, but since it is a substantial part (roughly 100 single-spaced
> > 
> > ```
> > 
> > pages out of 600 at the moment) of the book I am writing, I don’t  
> > think I can post it, though I might possibly be able to provide  
> > drafts to individuals who have the background and who would be  
> > prepared to offer helpful comments. All the same, I think it not  
> > unreasonable to post a small section that is a sort of introduction  
> > to the language part. Here is a PCT “just So Story” about how a baby  
> > produces and mother learns a first word.
> > 
> > ```
> > The working title of the book is "* Powers of Perceptual Control: a
> > 
> > ```
> > 
> > PCT enquiry into Language, Culture, Power, and Politics\* ", pun  
> > intended. My LCS IV chapter is more on the mechanism of protocols,  
> > and not really relevant to the question. However, some of my old  
> > Layered Protocol papers and chapters might be relevant, since LPT is  
> > just a special case of PCT invented before I knew of PCT. A few are  
> > on researchgate.
> > 
> > ```
> > Martin
> > 
> > ```

---

<div class="post-metadata">

**Author:** ![MartinT](http://discourse.iapct.org/letter_avatar_proxy/v4/letter/m/ee7513/32.png) [@MartinT](http://discourse.iapct.org/u/MartinT)\
**Post date:** [December 17, 2017, 5:17am UTC](http://discourse.iapct.org/t/natural-language-search/7509/9 "2017-12-17T05:17:49Z")

</div>

[Martin Taylor 2017.12.16.23.06]

> [From Bruce Nevin (2017.12.16.18:00 ET)]

> ```
> Martin Taylor
> 
> ```
> 
> 2017.12.15.13.14 –

> 

> ```
> It might be useful to have
> 
> ```
> 
> your definition of ‘protocol’, Martin.

```
Sorry. I take too much for granted. Since we had had a thread about

```

protocols and layered protocols some time ago, and since I’ve been  
writing about them and discussing them for over 30 years, I forgot  
that many people still won’t know much, if anything, about them.

```
Very simply, my definition of a protocol is that it is a procedure

```

whereby two people can discover what they can do to reduce error in  
a perception being controlled by another, and in the process reduce  
error in their own that is influenced by the actions of the other.  
As an example. Jane disturbs some perception John controls (e.g.his  
perception of Jane’s contentment) by saying “I’m cold. Could you  
shut the window”. I call that statement a “display” by Jane. It  
disturbs John’s perception of her contentment. He acts to bring that  
perception nearer its reference level by closing the window. The  
error in Jane’s temperature perception, and thus her comfort level  
perception, is thus reduced. She feels warmer.

```
I know that's much to terse to be much help in describing the usual

```

form of protocols. The main process is the discovery by both parties  
of what perception(s) each other is acting to control by influencing  
(the initiator by disturbing) a perception controlled by the other.  
Frequently the controlled perception is of the value(s) of  
perception(s) held by the other, in a type of protocol I label  
“Inform” or “Query”. I am the “continuer” in a “Query” protocol at  
this moment. I am also an initiator in an “Inform”, because I am  
doing more than just answering. I am acting to control a perception  
of your understanding of “protocol” that was disturbed by your  
query.

```
I don't know how many major types of protocol should be

```

distinguished. The process, however, is fairly well defined, at  
least for dyadic (between two persons) protocols. You can get a lot  
more detail in the two papers on the topic in the PCT Special Issue  
of Int ,J, Human-Computer Studies (Farrell, Hollands, Taylor and  
Gamble, 1998 and Taylor., Farrell, and Hollands, 1998). I thought I  
had uploaded copies of those papers to Researchgate, but apparently  
I didn’t.

```
Hope this helps, and is not too dense.

Martin

```

> **···**
>
> On Fri, Dec 15, 2017 at 1:49 PM, Martin  
> Taylor [mmt-csg@mmtaylor.net](mailto:mmt-csg@mmtaylor.net)  
> wrote:
> 
> > [Martin Taylor 2017.12.15.13.14]
> 
> > > [From Rupert Young (2017.12.12 09.30)]
> 
> > > ```
> > > ...Do we have anything that puts language (textual)
> > > 
> > > ```
> > > 
> > > in the PCT framework? …
> 
> > > 
> > 
> > ```
> > I do, but since it is a substantial part (roughly 100
> > 
> > ```
> > 
> > single-spaced pages out of 600 at the moment) of the book  
> > I am writing, I don’t think I can post it, though I might  
> > possibly be able to provide drafts to individuals who have  
> > the background and who would be prepared to offer helpful  
> > comments. All the same, I think it not unreasonable to  
> > post a small section that is a sort of introduction to the  
> > language part. Here is a PCT “just So Story” about how a  
> > baby produces and mother learns a first word.
> > 
> > ```
> > The working title of the book is "* Powers of Perceptual
> > 
> > ```
> > 
> > Control: a PCT enquiry into Language, Culture, Power,  
> > and Politics\* ", pun intended. My LCS IV chapter is  
> > more on the mechanism of protocols, and not really  
> > relevant to the question. However, some of my old Layered  
> > Protocol papers and chapters might be relevant, since LPT  
> > is just a special case of PCT invented before I knew of  
> > PCT. A few are on researchgate.
> > 
> > ```
> > Martin
> > 
> > ```

---

<div class="post-metadata">

**Author:** ![bnhpct](http://discourse.iapct.org/user_avatar/discourse.iapct.org/bnhpct/32/34_2.png) [@bnhpct](http://discourse.iapct.org/u/bnhpct)\
**Post date:** [December 17, 2017, 2:46pm UTC](http://discourse.iapct.org/t/natural-language-search/7509/10 "2017-12-17T14:46:22Z")

</div>

[From Bruce Nevin (2017.12.17.09:40 ET)]

Martin Taylor 2017.12.16.23.06 –

So a protocol (in this sense) is a perception that is controlled collectively by all parties in the communication. It is a Sequence perception about the process of communication, and probably sometimes a Program perception. (The distinction between Sequence and Program is evident logically but perhaps not always so crisp in wetware.) As such, its inputs, in the canonical model, are Category perceptions. The mother-infant interaction in your just-so story posits the emergence of a collectively-controlled perception (the protocol) by ‘co-reorganization’ in the course of a social interaction.

> **···**
>
> On Sun, Dec 17, 2017 at 12:17 AM, Martin Taylor \< mmt-csg@mmt  
> [aylor.net](http://aylor.net)\> wrote:
> 
> > [Martin Taylor 2017.12.16.23.06]
> 
> > > [From Bruce Nevin (2017.12.16.18:00 ET)]
> 
> > > ```
> > > Martin Taylor
> > > 
> > > ```
> > > 
> > > 2017.12.15.13.14 –
> 
> > >
> 
> > > ```
> > > It might be useful to have
> > > 
> > > ```
> > > 
> > > your definition of ‘protocol’, Martin.
> 
> > ```
> > Sorry. I take too much for granted. Since we had had a thread about
> > 
> > ```
> > 
> > protocols and layered protocols some time ago, and since I’ve been  
> > writing about them and discussing them for over 30 years, I forgot  
> > that many people still won’t know much, if anything, about them.
> > 
> > ```
> > Very simply, my definition of a protocol is that it is a procedure
> > 
> > ```
> > 
> > whereby two people can discover what they can do to reduce error in  
> > a perception being controlled by another, and in the process reduce  
> > error in their own that is influenced by the actions of the other.  
> > As an example. Jane disturbs some perception John controls (e.g.his  
> > perception of Jane’s contentment) by saying “I’m cold. Could you  
> > shut the window”. I call that statement a “display” by Jane. It  
> > disturbs John’s perception of her contentment. He acts to bring that  
> > perception nearer its reference level by closing the window. The  
> > error in Jane’s temperature perception, and thus her comfort level  
> > perception, is thus reduced. She feels warmer.
> > 
> > ```
> > I know that's much to terse to be much help in describing the usual
> > 
> > ```
> > 
> > form of protocols. The main process is the discovery by both parties  
> > of what perception(s) each other is acting to control by influencing  
> > (the initiator by disturbing) a perception controlled by the other.  
> > Frequently the controlled perception is of the value(s) of  
> > perception(s) held by the other, in a type of protocol I label  
> > “Inform” or “Query”. I am the “continuer” in a “Query” protocol at  
> > this moment. I am also an initiator in an “Inform”, because I am  
> > doing more than just answering. I am acting to control a perception  
> > of your understanding of “protocol” that was disturbed by your  
> > query.
> > 
> > ```
> > I don't know how many major types of protocol should be
> > 
> > ```
> > 
> > distinguished. The process, however, is fairly well defined, at  
> > least for dyadic (between two persons) protocols. You can get a lot  
> > more detail in the two papers on the topic in the PCT Special Issue  
> > of Int ,J, Human-Computer Studies (Farrell, Hollands, Taylor and  
> > Gamble, 1998 and Taylor., Farrell, and Hollands, 1998). I thought I  
> > had uploaded copies of those papers to Researchgate, but apparently  
> > I didn’t.
> > 
> > ```
> > Hope this helps, and is not too dense.
> > 
> > Martin
> > 
> > ```
> > 
> > >
> 
> > > ```
> > > In the usual definitions,
> > > 
> > > ```
> > > 
> > > a protocol is (or specifies) constraints to which a  
> > > communicative interaction must conform in order to be  
> > > perceived as such. Presupposed metacommunication, if you  
> > > will. Here are two definitions (for diplomacy and for  
> > > computer intercommunication, respectively) from the Random  
> > > House Dictionary:
> 
> > >
> 
> > > ```
> > > 1.
> > > 
> > > ```
> > > 
> > > the customs and regulations dealing with diplomatic  
> > > formality, precedence, and etiquette.
> 
> > > ```
> > > [...]
> > > 
> > > ```
> 
> > > ```
> > > 7.
> > > 
> > > ```
> > > 
> > > Computers. a set of rules governing the format of messages  
> > > that are exchanged between computers.
> 
> > >
> 
> > > ```
> > > Thus, the Internet
> > > 
> > > ```
> > > 
> > > Protocol specifies a number of things (such as IP  
> > > addresses) to which communications from one computer  
> > > across the Internet to another must comply.
> 
> > >
> 
> > > ```
> > > You refer to the baby's
> > > 
> > > ```
> > > 
> > > initiating protocol and the mother’s continuation  
> > > protocol. The word here seems to refer to parts of an  
> > > instance of communication rather than to the whole. Maybe  
> > > this is analogous to the definitions 2-5 elided above.  
> > > They are:
> 
> > >
> 
> > > ```
> > > 2.
> > > 
> > > ```
> > > 
> > > an original draft, minute, or record from which a  
> > > document, esp. a treaty, is prepared.
> 
> > > ```
> > > 3.
> > > 
> > > ```
> > > 
> > > a supplementary international agreement.
> 
> > > ```
> > > 4.
> > > 
> > > ```
> > > 
> > > an agreement between states.
> 
> > > ```
> > > 5.
> > > 
> > > ```
> > > 
> > > an annex to a treaty giving data relating to it.
> 
> > >
> 
> > > ```
> > > In these cases, it's not
> > > 
> > > ```
> > > 
> > > synecdoche–the part taken to represent the whole–but  
> > > rather the word for the whole (‘protocol’) is used to  
> > > refer to a part. The analogy breaks down because these are  
> > > parts of a particular instance of protocol-compliant  
> > > communication (negotiation of a treaty, etc.), and your  
> > > terms initiating protocol and continuation protocol are at  
> > > a higher level of generalization. An initiating protocol  
> > > is any observed behavior by party A that is [intended to  
> > > be] perceived by party B as initiating an an exchange  
> > > compliant with a known protocol; and so on.
> 
> > >
> 
> > > ```
> > > There are two other
> > > 
> > > ```
> > > 
> > > dictionary definitions which I believe are a bit too  
> > > divergent to be relevant:
> 
> > >
> 
> > > ```
> > > 6.
> > > 
> > > ```
> > > 
> > > Med. the plan for carrying out a scientific study or a  
> > > patient’s treatment regimen.
> 
> > > ```
> > > 8.
> > > 
> > > ```
> > > 
> > > Also called protocol statement, protocol sentence,  
> > > protocol proposition. Philos. a statement reporting an  
> > > observation or experience in the most fundamental terms  
> > > without interpretation: sometimes taken as the basis of  
> > > empirical verification, as of scientific laws.
> 
> > >
> 
> > > > /Bruce
> 
> > >
> 
> > 
> 
> ```
> On Fri, Dec 15, 2017 at 1:49 PM, Martin
> 
> ```
> 
> Taylor [mmt-csg@mmtaylor.net](mailto:mmt-csg@mmtaylor.net)  
> wrote:
> 
> > [Martin Taylor 2017.12.15.13.14]
> 
> > > [From Rupert Young (2017.12.12 09.30)]
> 
> > > ```
> > > ...Do we have anything that puts language (textual)
> > > 
> > > ```
> > > 
> > > in the PCT framework? …
> 
> > > 
> > 
> > ```
> > I do, but since it is a substantial part (roughly 100
> > 
> > ```
> > 
> > single-spaced pages out of 600 at the moment) of the book  
> > I am writing, I don’t think I can post it, though I might  
> > possibly be able to provide drafts to individuals who have  
> > the background and who would be prepared to offer helpful  
> > comments. All the same, I think it not unreasonable to  
> > post a small section that is a sort of introduction to the  
> > language part. Here is a PCT “just So Story” about how a  
> > baby produces and mother learns a first word.
> > 
> > ```
> > The working title of the book is "* Powers of Perceptual
> > 
> > ```
> > 
> > Control: a PCT enquiry into Language, Culture, Power,  
> > and Politics\* ", pun intended. My LCS IV chapter is  
> > more on the mechanism of protocols, and not really  
> > relevant to the question. However, some of my old Layered  
> > Protocol papers and chapters might be relevant, since LPT  
> > is just a special case of PCT invented before I knew of  
> > PCT. A few are on researchgate.
> > 
> > ```
> > Martin
> > 
> > ```

---

<div class="post-metadata">

**Author:** ![rsmarken](http://discourse.iapct.org/user_avatar/discourse.iapct.org/rsmarken/32/3553_2.png) [@rsmarken](http://discourse.iapct.org/u/rsmarken)\
**Post date:** [December 17, 2017, 6:42pm UTC](http://discourse.iapct.org/t/natural-language-search/7509/11 "2017-12-17T18:42:42Z")

</div>

[From Rick Marken (2017.12.17.1040)]

> **···**
>
> > Bruce Nevin (2017.12.17.09:40 ET)
> 
> > 
> 
> > BN: So a protocol (in this sense) is a perception that is controlled collectively by all parties in the communication.
> 
> RM: Right. A “protocol” is a theoretical concept (defined in terms of a hybrid of PCT and reinforcement theory) in search of a phenomenon!Â Â
> 
> Â
> 
> > It is a Sequence perception about the process of communication, and probably sometimes a Program perception. (The distinction between Sequence and Program is evident logically but perhaps not always so crisp in wetware.) As such, its inputs, in the canonical model, are Category perceptions. The mother-infant interaction in your just-so story posits the emergence of a collectively-controlled perception (the protocol) by ‘co-reorganization’ in the course of a social interaction.
> 
> –
> 
> On Sun, Dec 17, 2017 at 12:17 AM, Martin Taylor \< mmt-csg@mmt  
> [aylor.net](http://aylor.net)\> wrote:
> 
> > [Martin Taylor 2017.12.16.23.06]
> 
> > > [From Bruce Nevin (2017.12.16.18:00 ET)]
> 
> > > ```
> > > Martin Taylor
> > > 
> > > ```
> > > 
> > > 2017.12.15.13.14 –
> 
> > >
> 
> > > ```
> > > It might be useful to have
> > > 
> > > ```
> > > 
> > > your definition of ‘protocol’, Martin.
> 
> > ```
> > Sorry. I take too much for granted. Since we had had a thread about
> > 
> > ```
> > 
> > protocols and layered protocols some time ago, and since I’ve been  
> > writing about them and discussing them for over 30 years, I forgot  
> > that many people still won’t know much, if anything, about them.
> > 
> > ```
> > Very simply, my definition of a protocol is that it is a procedure
> > 
> > ```
> > 
> > whereby two people can discover what they can do to reduce error in  
> > a perception being controlled by another, and in the process reduce  
> > error in their own that is influenced by the actions of the other.  
> > As an example. Jane disturbs some perception John controls (e.g.his  
> > perception of Jane’s contentment) by saying “I’m cold. Could you  
> > shut the window”. I call that statement a “display” by Jane. It  
> > disturbs John’s perception of her contentment. He acts to bring that  
> > perception nearer its reference level by closing the window. The  
> > error in Jane’s temperature perception, and thus her comfort level  
> > perception, is thus reduced. She feels warmer.
> > 
> > ```
> > I know that's much to terse to be much help in describing the usual
> > 
> > ```
> > 
> > form of protocols. The main process is the discovery by both parties  
> > of what perception(s) each other is acting to control by influencing  
> > (the initiator by disturbing) a perception controlled by the other.  
> > Frequently the controlled perception is of the value(s) of  
> > perception(s) held by the other, in a type of protocol I label  
> > “Inform” or “Query”. I am the “continuer” in a “Query” protocol at  
> > this moment. I am also an initiator in an “Inform”, because I am  
> > doing more than just answering. I am acting to control a perception  
> > of your understanding of “protocol” that was disturbed by your  
> > query.
> > 
> > ```
> > I don't know how many major types of protocol should be
> > 
> > ```
> > 
> > distinguished. The process, however, is fairly well defined, at  
> > least for dyadic (between two persons) protocols. You can get a lot  
> > more detail in the two papers on the topic in the PCT Special Issue  
> > of Int ,J, Human-Computer Studies (Farrell, Hollands, Taylor and  
> > Gamble, 1998 and Taylor., Farrell, and Hollands, 1998). I thought I  
> > had uploaded copies of those papers to Researchgate, but apparently  
> > I didn’t.
> > 
> > ```
> > Hope this helps, and is not too dense.
> > 
> > Martin
> > 
> > ```
> > 
> > >
> 
> > > ```
> > > In the usual definitions,
> > > 
> > > ```
> > > 
> > > a protocol is (or specifies) constraints to which a  
> > > communicative interaction must conform in order to be  
> > > perceived as such. Presupposed metacommunication, if you  
> > > will. Here are two definitions (for diplomacy and for  
> > > computer intercommunication, respectively) from the Random  
> > > House Dictionary:
> 
> > >
> 
> > > ```
> > > 1.
> > > 
> > > ```
> > > 
> > > the customs and regulations dealing with diplomatic  
> > > formality, precedence, and etiquette.
> 
> > > ```
> > > [...]
> > > 
> > > ```
> 
> > > ```
> > > 7.
> > > 
> > > ```
> > > 
> > > Computers. a set of rules governing the format of messages  
> > > that are exchanged between computers.
> 
> > >
> 
> > > ```
> > > Thus, the Internet
> > > 
> > > ```
> > > 
> > > Protocol specifies a number of things (such as IP  
> > > addresses) to which communications from one computer  
> > > across the Internet to another must comply.
> 
> > >
> 
> > > ```
> > > You refer to the baby's
> > > 
> > > ```
> > > 
> > > initiating protocol and the mother’s continuation  
> > > protocol. The word here seems to refer to parts of an  
> > > instance of communication rather than to the whole. Maybe  
> > > this is analogous to the definitions 2-5 elided above.  
> > > They are:
> 
> > >
> 
> > > ```
> > > 2.
> > > 
> > > ```
> > > 
> > > an original draft, minute, or record from which a  
> > > document, esp. a treaty, is prepared.
> 
> > > ```
> > > 3.
> > > 
> > > ```
> > > 
> > > a supplementary international agreement.
> 
> > > ```
> > > 4.
> > > 
> > > ```
> > > 
> > > an agreement between states.
> 
> > > ```
> > > 5.
> > > 
> > > ```
> > > 
> > > an annex to a treaty giving data relating to it.
> 
> > >
> 
> > > ```
> > > In these cases, it's not
> > > 
> > > ```
> > > 
> > > synecdoche–the part taken to represent the whole–but  
> > > rather the word for the whole (‘protocol’) is used to  
> > > refer to a part. The analogy breaks down because these are  
> > > parts of a particular instance of protocol-compliant  
> > > communication (negotiation of a treaty, etc.), and your  
> > > terms initiating protocol and continuation protocol are at  
> > > a higher level of generalization. An initiating protocol  
> > > is any observed behavior by party A that is [intended to  
> > > be] perceived by party B as initiating an an exchange  
> > > compliant with a known protocol; and so on.
> 
> > >
> 
> > > ```
> > > There are two other
> > > 
> > > ```
> > > 
> > > dictionary definitions which I believe are a bit too  
> > > divergent to be relevant:
> 
> > >
> 
> > > ```
> > > 6.
> > > 
> > > ```
> > > 
> > > Med. the plan for carrying out a scientific study or a  
> > > patient’s treatment regimen.
> 
> > > ```
> > > 8.
> > > 
> > > ```
> > > 
> > > Also called protocol statement, protocol sentence,  
> > > protocol proposition. Philos. a statement reporting an  
> > > observation or experience in the most fundamental terms  
> > > without interpretation: sometimes taken as the basis of  
> > > empirical verification, as of scientific laws.
> 
> > >
> 
> > > > /Bruce
> 
> > >
> 
> > 
> 
> ```
> On Fri, Dec 15, 2017 at 1:49 PM, Martin
> 
> ```
> 
> Taylor [mmt-csg@mmtaylor.net](mailto:mmt-csg@mmtaylor.net)  
> wrote:
> 
> > [Martin Taylor 2017.12.15.13.14]
> 
> > > [From Rupert Young (2017.12.12 09.30)]
> 
> > > ```
> > > ...Do we have anything that puts language (textual)
> > > 
> > > ```
> > > 
> > > in the PCT framework? …
> 
> > > 
> > 
> > ```
> > I do, but since it is a substantial part (roughly 100
> > 
> > ```
> > 
> > single-spaced pages out of 600 at the moment) of the book  
> > I am writing, I don’t think I can post it, though I might  
> > possibly be able to provide drafts to individuals who have  
> > the background and who would be prepared to offer helpful  
> > comments. All the same, I think it not unreasonable to  
> > post a small section that is a sort of introduction to the  
> > language part. Here is a PCT “just So Story” about how a  
> > baby produces and mother learns a first word.
> > 
> > ```
> > The working title of the book is "* Powers of Perceptual
> > 
> > ```
> > 
> > Control: a PCT enquiry into Language, Culture, Power,  
> > and Politics\* ", pun intended. My LCS IV chapter is  
> > more on the mechanism of protocols, and not really  
> > relevant to the question. However, some of my old Layered  
> > Protocol papers and chapters might be relevant, since LPT  
> > is just a special case of PCT invented before I knew of  
> > PCT. A few are on researchgate.
> > 
> > ```
> > Martin
> > 
> > ```
> 
> Richard S. MarkenÂ
> 
> "Perfection is achieved not when you have nothing more to add, but when you  
> have nothing left to take away.â?  
> Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â --Antoine de Saint-Exupery

---

<div class="post-metadata">

**Author:** ![MartinT](http://discourse.iapct.org/letter_avatar_proxy/v4/letter/m/ee7513/32.png) [@MartinT](http://discourse.iapct.org/u/MartinT)\
**Post date:** [December 17, 2017, 7:26pm UTC](http://discourse.iapct.org/t/natural-language-search/7509/12 "2017-12-17T19:26:16Z")

</div>

[Martin Taylor 2017.12.17.11.17]

> [From Bruce Nevin (2017.12.17.09:40 ET)]

> 

> ```
> Martin Taylor
> 
> ```
> 
> 2017.12.16.23.06 –

> 

> ```
> So a protocol (in this sense) is a perception that is
> 
> ```
> 
> controlled collectively by all parties in the communication.

```
I was afraid that I was much too terse in my description. The full

```

description of the possible progress of a protocol involves as many  
as nineteen different controlled perceptions. Of these nineteen,  
there are nine that are common to all protocols in each participant  
in a dialogue, though in most instances several, even as many as  
eight, are undisturbed and therefore produce no overt action. The  
nineteenth is the perception by the “sender of a message” (a.k.a the  
“initiator of a protocol”) of what the “sender of the message” wants  
the recipient (a.k.a “the continuer of a protocol”) to perceive.  
That perception is controlled by the sender; the others are all in  
support of the sender’s ability to control that one. The “meaning”  
of the message (Ivan’s “hungry-word”, for example) probably is  
collectively controlled, not just by the dialogue partners, though  
that is possible, but by the larger linguistic community into which  
the child emerges. The “meaning” may be different between the sender  
and the recipient, but usually the difference is slight or  
non-existent.

```
Different reference values for the various members of the supporting

```

group of perceptions controlled in the course of a protocol  
determine whether the intentions of the initiator and the recipient  
are cooperative, collaborative, deceitful, and so forth. In actual  
dialogue, protocols have multiple levels of support, just as do  
control units in the hierarchy. A dialogue can be cooperative at one  
level and deceitful by one or both of the parties at another level.

```
Nineteen controlled perceptions may sound like unnecessary bells and

```

whistles, not to mention bows and ribbons. But every one of them  
seems to be required to account for the general case. When we were  
working on that aspect of protocols, we anticipated the possibility  
of an infinite recursion, which would have resulted in 2\*3n +1  
possible controlled perceptions, but when we examined both real and  
hypothetical dialogues, we could find no instances for n\>2,  
whereas we could find instances where n=2. And if you want to go  
further, we could find only about 45 plausible different kinds of  
actions for controlling these 18 perceptions. All of which doesn’t  
mean that we (Taylor Farrell and Hollands 1998) were correct, but it  
does mean that unless someone finds a simpler way to describe  
dialogue interactions, the progress of a protocol is determined to a  
large extent by the ever-changing values of what each partner  
perceives the other to perceive while the main message is being  
passed.

```
Apart from all the papers and chapters I and my colleagues wrote on

```

Layered Protocols before I learned of PCT, the book “Powers of  
Perceptual Control” has a full chapter describing protocols as such,  
and another chapter mostly devoted to how protocols develop and are  
collectively controlled as perceptions in a community or culture.  
That chapter also discusses how they differ from a PCT viewpoint  
from rituals, morals, and laws. I can make these chapters (all  
drafts possibly subject to much revision) available to you – as  
someone much concerned with the use of language – for comment, but  
I won’t post them more widely.

```
I think a protocol is an emergent property in the same way as is

```

control, stiffness, or tensegrity. Each has a minimum structure that  
is needed before a protocol can exist, just as there is for control.  
Stiffness requires at least two control units, tensegrity probably  
nine, and protocol possibly nineteen. What is not needed, though it  
may exist and may be collectively controlled, is a perception in  
either party of the form of the protocol (I use a specific example  
of collective control of a protocol form for various purposes in  
Chapters 17-19).

> ```
> It is a Sequence perception about the process of
> 
> ```
> 
> communication, and probably sometimes a Program perception.  
> (The distinction between Sequence and Program is evident  
> logically but perhaps not always so crisp in wetware.) As  
> such, its inputs, in the canonical model, are Category  
> perceptions. The mother-infant interaction in your just-so  
> story posits the emergence of a collectively-controlled  
> perception (the protocol) by ‘co-reorganization’ in the course  
> of a social interaction.

```
I don't think there's necessarily a collectively controlled

```

perception other than that an arbitrary pattern of sound and  
movement begins to have a “meaning” (another emergent property that  
requires a minimum supporting structure). Baby Ivan controls only  
one of the nine that an adult might (“that mother has correctly  
interpreted the message”, the message being “feed me” or “I’m  
hungry” or something similar for which Cora’s action in feeding him  
reduces the error).

```
That said, I also argue that many specific forms of protocol are

```

collectively controlled, just as are the meanings of some  
sound-gesture patterns that in the distant pre-history of a language  
were arbitrary. In the extreme, these collectively controlled forms  
of protocols can sometimes merge into the forms of rituals. The  
difference with rituals is that in rituals the defining property is  
that the actions of the ritual may be publicly perceived and  
controlled. An improperly executed ritual does not serve its  
purpose. For example, a marriage service performed by a  
non-authorized person or in which the potential spouses say “No I  
won’t” rather than “I do” will not result in the partners being  
perceived by most people as married. The marriage service is a  
ritual, though it may look like a triadic protocol because of the  
three-way interaction among the partners and the officiating  
person…

```
Martin

```

> **···**
>
> On Sun, Dec 17, 2017 at 12:17 AM,  
> Martin Taylor \<mmt-csg@mmt [aylor.net](http://aylor.net)\>  
> wrote:
> 
> > [Martin Taylor 2017.12.16.23.06]
> 
> > > ```
> > > [From Bruce Nevin (2017.12.16.18:00
> > > 
> > > ```
> > > 
> > > ET)]
> 
> > > ```
> > > Martin Taylor
> > > 
> > > ```
> > > 
> > > 2017.12.15.13.14 –
> 
> > >
> 
> > > ```
> > > It might be
> > > 
> > > ```
> > > 
> > > useful to have your definition of ‘protocol’,  
> > > Martin.
> 
> > ```
> > Sorry. I take too much for granted. Since we had
> > 
> > ```
> > 
> > had a thread about protocols and layered protocols some  
> > time ago, and since I’ve been writing about them and  
> > discussing them for over 30 years, I forgot that many  
> > people still won’t know much, if anything, about them.
> > 
> > ```
> > Very simply, my definition of a protocol is that it is a
> > 
> > ```
> > 
> > procedure whereby two people can discover what they can do  
> > to reduce error in a perception being controlled by  
> > another, and in the process reduce error in their own that  
> > is influenced by the actions of the other. As an example.  
> > Jane disturbs some perception John controls (e.g.his  
> > perception of Jane’s contentment) by saying “I’m cold.  
> > Could you shut the window”. I call that statement a  
> > “display” by Jane. It disturbs John’s perception of her  
> > contentment. He acts to bring that perception nearer its  
> > reference level by closing the window. The error in Jane’s  
> > temperature perception, and thus her comfort level  
> > perception, is thus reduced. She feels warmer.
> > 
> > ```
> > I know that's much to terse to be much help in describing
> > 
> > ```
> > 
> > the usual form of protocols. The main process is the  
> > discovery by both parties of what perception(s) each other  
> > is acting to control by influencing (the initiator by  
> > disturbing) a perception controlled by the other.  
> > Frequently the controlled perception is of the value(s) of  
> > perception(s) held by the other, in a type of protocol I  
> > label “Inform” or “Query”. I am the “continuer” in a  
> > “Query” protocol at this moment. I am also an initiator in  
> > an “Inform”, because I am doing more than just answering.  
> > I am acting to control a perception of your understanding  
> > of “protocol” that was disturbed by your query.
> > 
> > ```
> > I don't know how many major types of protocol should be
> > 
> > ```
> > 
> > distinguished. The process, however, is fairly well  
> > defined, at least for dyadic (between two persons)  
> > protocols. You can get a lot more detail in the two papers  
> > on the topic in the PCT Special Issue of Int ,J,  
> > Human-Computer Studies (Farrell, Hollands, Taylor and  
> > Gamble, 1998 and Taylor., Farrell, and Hollands, 1998). I  
> > thought I had uploaded copies of those papers to  
> > Researchgate, but apparently I didn’t.
> > 
> > ```
> > Hope this helps, and is not too dense.
> > 
> > Martin
> > 
> > ```
> > 
> > >
> 
> > > ```
> > > In the usual
> > > 
> > > ```
> > > 
> > > definitions, a protocol is (or specifies)  
> > > constraints to which a communicative  
> > > interaction must conform in order to be  
> > > perceived as such. Presupposed  
> > > metacommunication, if you will. Here are two  
> > > definitions (for diplomacy and for computer  
> > > intercommunication, respectively) from the  
> > > Random House Dictionary:
> 
> > >
> 
> > > ```
> > > 1.
> > > 
> > > ```
> > > 
> > > the customs and regulations dealing with  
> > > diplomatic formality, precedence, and  
> > > etiquette.
> 
> > > ```
> > > [...]
> > > 
> > > ```
> 
> > > ```
> > > 7.
> > > 
> > > ```
> > > 
> > > Computers. a set of rules governing the format  
> > > of messages that are exchanged between  
> > > computers.
> 
> > >
> 
> > > ```
> > > Thus, the
> > > 
> > > ```
> > > 
> > > Internet Protocol specifies a number of things  
> > > (such as IP addresses) to which communications  
> > > from one computer across the Internet to  
> > > another must comply.
> 
> > >
> 
> > > ```
> > > You refer to
> > > 
> > > ```
> > > 
> > > the baby’s initiating protocol and the  
> > > mother’s continuation protocol. The word here  
> > > seems to refer to parts of an instance of  
> > > communication rather than to the whole. Maybe  
> > > this is analogous to the definitions 2-5  
> > > elided above. They are:
> 
> > >
> 
> > > ```
> > > 2.
> > > 
> > > ```
> > > 
> > > an original draft, minute, or record from  
> > > which a document, esp. a treaty, is prepared.
> 
> > > ```
> > > 3.
> > > 
> > > ```
> > > 
> > > a supplementary international agreement.
> 
> > > ```
> > > 4.
> > > 
> > > ```
> > > 
> > > an agreement between states.
> 
> > > ```
> > > 5.
> > > 
> > > ```
> > > 
> > > an annex to a treaty giving data relating to  
> > > it.
> 
> > >
> 
> > > ```
> > > In these
> > > 
> > > ```
> > > 
> > > cases, it’s not synecdoche–the part taken to  
> > > represent the whole–but rather the word for  
> > > the whole (‘protocol’) is used to refer to a  
> > > part. The analogy breaks down because these  
> > > are parts of a particular instance of  
> > > protocol-compliant communication (negotiation  
> > > of a treaty, etc.), and your terms initiating  
> > > protocol and continuation protocol are at a  
> > > higher level of generalization. An initiating  
> > > protocol is any observed behavior by party A  
> > > that is [intended to be] perceived by party B  
> > > as initiating an an exchange compliant with a  
> > > known protocol; and so on.
> 
> > >
> 
> > > ```
> > > There are two
> > > 
> > > ```
> > > 
> > > other dictionary definitions which I believe  
> > > are a bit too divergent to be relevant:
> 
> > >
> 
> > > ```
> > > 6.
> > > 
> > > ```
> > > 
> > > Med. the plan for carrying out a scientific  
> > > study or a patient’s treatment regimen.
> 
> > > ```
> > > 8.
> > > 
> > > ```
> > > 
> > > Also called protocol statement, protocol  
> > > sentence, protocol proposition. Philos. a  
> > > statement reporting an observation or  
> > > experience in the most fundamental terms  
> > > without interpretation: sometimes taken as the  
> > > basis of empirical verification, as of  
> > > scientific laws.
> 
> > >
> 
> > > > /Bruce
> 
> > >
> 
> > 
> 
> ```
> On Fri, Dec 15, 2017 at
> 
> ```
> 
> 1:49 PM, Martin Taylor [mmt-csg@mmtaylor.net](mailto:mmt-csg@mmtaylor.net)  
> wrote:
> 
> > [Martin Taylor 2017.12.15.13.14]
> 
> > > [From Rupert Young (2017.12.12 09.30)]
> 
> > > ```
> > > ...Do we have anything that puts language
> > > 
> > > ```
> > > 
> > > (textual) in the PCT framework? …
> 
> > > 
> > 
> > ```
> > I do, but since it is a substantial part
> > 
> > ```
> > 
> > (roughly 100 single-spaced pages out of 600 at  
> > the moment) of the book I am writing, I don’t  
> > think I can post it, though I might possibly  
> > be able to provide drafts to individuals who  
> > have the background and who would be prepared  
> > to offer helpful comments. All the same, I  
> > think it not unreasonable to post a small  
> > section that is a sort of introduction to the  
> > language part. Here is a PCT “just So Story”  
> > about how a baby produces and mother learns a  
> > first word.
> > 
> > ```
> > The working title of the book is "* Powers of
> > 
> > ```
> > 
> > Perceptual Control: a PCT enquiry into  
> > Language, Culture, Power, and Politics\* ",  
> > pun intended. My LCS IV chapter is more on the  
> > mechanism of protocols, and not really  
> > relevant to the question. However, some of my  
> > old Layered Protocol papers and chapters might  
> > be relevant, since LPT is just a special case  
> > of PCT invented before I knew of PCT. A few  
> > are on researchgate.
> > 
> > ```
> > Martin
> > 
> > ```

---

<div class="post-metadata">

**Author:** ![\_Boris\_Hartman](http://discourse.iapct.org/letter_avatar_proxy/v4/letter/_/e8c25b/32.png) [@\_Boris\_Hartman](http://discourse.iapct.org/u/_Boris_Hartman)\
**Post date:** [December 18, 2017, 6:57pm UTC](http://discourse.iapct.org/t/natural-language-search/7509/13 "2017-12-18T18:57:55Z")

</div>

Martin,

After a long time I’m glad to »hear« pure PCT. Superb explanation of PCT…

Boris

> **···**
>
> **From:** Martin Taylor [[mailto:mmt-csg@mmtaylor.net](mailto:mmt-csg@mmtaylor.net)]  
> **Sent:** Sunday, December 17, 2017 6:18 AM  
> **To:** csgnet@lists.illinois.edu  
> **Subject:** Re: Natural language search
> 
> [Martin Taylor 2017.12.16.23.06]
> 
> [From Bruce Nevin (2017.12.16.18:00 ET)]
> 
> Martin Taylor 2017.12.15.13.14 –
> 
> It might be useful to have your definition of ‘protocol’, Martin.
> 
> Sorry. I take too much for granted. Since we had had a thread about protocols and layered protocols some time ago, and since I’ve been writing about them and discussing them for over 30 years, I forgot that many people still won’t know much, if anything, about them.
> 
> Very simply, my definition of a protocol is that it is a procedure whereby two people can discover what they can do to reduce error in a perception being controlled by another, and in the process reduce error in their own that is influenced by the actions of the other. As an example. Jane disturbs some perception John controls (e.g.his perception of Jane’s contentment) by saying “I’m cold. Could you shut the window”. I call that statement a “display” by Jane. It disturbs John’s perception of her contentment. He acts to bring that perception nearer its reference level by closing the window. The error in Jane’s temperature perception, and thus her comfort level perception, is thus reduced. She feels warmer.
> 
> I know that’s much to terse to be much help in describing the usual form of protocols. The main process is the discovery by both parties of what perception(s) each other is acting to control by influencing (the initiator by disturbing) a perception controlled by the other. Frequently the controlled perception is of the value(s) of perception(s) held by the other, in a type of protocol I label “Inform” or “Query”. I am the “continuer” in a “Query” protocol at this moment. I am also an initiator in an “Inform”, because I am doing more than just answering. I am acting to control a perception of your understanding of “protocol” that was disturbed by your query.
> 
> I don’t know how many major types of protocol should be distinguished. The process, however, is fairly well defined, at least for dyadic (between two persons) protocols. You can get a lot more detail in the two papers on the topic in the PCT Special Issue of Int ,J, Human-Computer Studies (Farrell, Hollands, Taylor and Gamble, 1998 and Taylor., Farrell, and Hollands, 1998). I thought I had uploaded copies of those papers to Researchgate, but apparently I didn’t.
> 
> Hope this helps, and is not too dense.
> 
> Martin
> 
> In the usual definitions, a protocol is (or specifies) constraints to which a communicative interaction must conform in order to be perceived as such. Presupposed metacommunication, if you will. Here are two definitions (for diplomacy and for computer intercommunication, respectively) from the Random House Dictionary:
> 
> ```
> 1. the customs and regulations dealing with diplomatic formality, precedence, and etiquette.
> 
> [...]
> 
> 7. Computers. a set of rules governing the format of messages that are exchanged between computers.
> 
> ```
> 
> Thus, the Internet Protocol specifies a number of things (such as IP addresses) to which communications from one computer across the Internet to another must comply.
> 
> You refer to the baby’s initiating protocol and the mother’s continuation protocol. The word here seems to refer to parts of an instance of communication rather than to the whole. Maybe this is analogous to the definitions 2-5 elided above. They are:
> 
> ```
> 2. an original draft, minute, or record from which a document, esp. a treaty, is prepared.
> 
> 3. a supplementary international agreement.
> 
> 4. an agreement between states.
> 
> 5. an annex to a treaty giving data relating to it.
> 
> ```
> 
> In these cases, it’s not synecdoche–the part taken to represent the whole–but rather the word for the whole (‘protocol’) is used to refer to a part. The analogy breaks down because these are parts of a particular instance of protocol-compliant communication (negotiation of a treaty, etc.), and your terms initiating protocol and continuation protocol are at a higher level of generalization. An initiating protocol is any observed behavior by party A that is [intended to be] perceived by party B as initiating an an exchange compliant with a known protocol; and so on.
> 
> There are two other dictionary definitions which I believe are a bit too divergent to be relevant:
> 
> ```
> 6. Med. the plan for carrying out a scientific study or a patient's treatment regimen.
> 
> 8. Also called protocol statement, protocol sentence, protocol proposition. Philos. a statement reporting an observation or experience in the most fundamental terms without interpretation: sometimes taken as the basis of empirical verification, as of scientific laws.
> 
> ```
> 
> > /Bruce
> 
> On Fri, Dec 15, 2017 at 1:49 PM, Martin Taylor [mmt-csg@mmtaylor.net](mailto:mmt-csg@mmtaylor.net) wrote:
> 
> [Martin Taylor 2017.12.15.13.14]
> 
> > [From Rupert Young (2017.12.12 09.30)]
> 
> > …Do we have anything that puts language (textual) in the PCT framework? …
> 
> I do, but since it is a substantial part (roughly 100 single-spaced pages out of 600 at the moment) of the book I am writing, I don’t think I can post it, though I might possibly be able to provide drafts to individuals who have the background and who would be prepared to offer helpful comments. All the same, I think it not unreasonable to post a small section that is a sort of introduction to the language part. Here is a PCT “just So Story” about how a baby produces and mother learns a first word.  
> The working title of the book is “_Powers of Perceptual Control: a PCT enquiry into Language, Culture, Power, and Politics_”, pun intended. My LCS IV chapter is more on the mechanism of protocols, and not really relevant to the question. However, some of my old Layered Protocol papers and chapters might be relevant, since LPT is just a special case of PCT invented before I knew of PCT. A few are on researchgate.
> 
> Martin

---

<div class="post-metadata">

**Author:** ![\_Boris\_Hartman](http://discourse.iapct.org/letter_avatar_proxy/v4/letter/_/e8c25b/32.png) [@\_Boris\_Hartman](http://discourse.iapct.org/u/_Boris_Hartman)\
**Post date:** [December 18, 2017, 7:41pm UTC](http://discourse.iapct.org/t/natural-language-search/7509/14 "2017-12-18T19:41:49Z")

</div>

Down

> **···**
>
> **From:** Richard Marken [[mailto:rsmarken@gmail.com](mailto:rsmarken@gmail.com)]  
> **Sent:** Sunday, December 17, 2017 7:43 PM  
> **To:** csgnet@lists.illinois.edu  
> **Subject:** Re: Natural language search
> 
> [From Rick Marken (2017.12.17.1040)]
> 
> > Bruce Nevin (2017.12.17.09:40 ET)
> 
> > BN: So a protocol (in this sense) is a perception that is controlled collectively by all parties in the communication.
> 
> RM: Right. A “protocol” is a theoretical concept (defined in terms of a hybrid of PCT and reinforcement theory) in search of a phenomenon!
> 
> HB : Whenever you Â»open your mouthÂ« you are outdoing yourself in stupidity.
> 
> RM earlier : In my rush to show that this is not the case I came up with what has to be the dumbest rebuttal of all time – outdoing even myself in stupidity;-)
> 
> HB : Martin gave a perfect description of PCT Â»protokolÂ«, which I think is perfect match to Kents’ Â»collective controlÂ«.
> 
> And the best thing is that is not theoretical concept at all. You can try it anywhere in any life situation
> 
> RM earlier : But if one’s goal is to understand the behavior of living organisms, I would also recommend comparing the behavior of these working models to that of real organisms. This is something that has been sorely lacking in these discussions.
> 
> HB : So why you don’t start with real organismsÂ and real life situations ???
> 
> Boris
> 
> > It is a Sequence perception about the process of communication, and probably sometimes a Program perception. (The distinction between Sequence and Program is evident logically but perhaps not always so crisp in wetware.) As such, its inputs, in the canonical model, are Category perceptions. The mother-infant interaction in your just-so story posits the emergence of a collectively-controlled perception (the protocol) by ‘co-reorganization’ in the course of a social interaction.
> 
> > On Sun, Dec 17, 2017 at 12:17 AM, Martin Taylor \<mmt-csg@mmt [aylor.net](http://aylor.net)\> wrote:
> 
> > [Martin Taylor 2017.12.16.23.06]
> 
> > 
> 
> > [From Bruce Nevin (2017.12.16.18:00 ET)]
> 
> > Martin Taylor 2017.12.15.13.14 –
> 
> > It might be useful to have your definition of ‘protocol’, Martin.
> 
> > Sorry. I take too much for granted. Since we had had a thread about protocols and layered protocols some time ago, and since I’ve been writing about them and discussing them for over 30 years, I forgot that many people still won’t know much, if anything, about them.
> > 
> > Very simply, my definition of a protocol is that it is a procedure whereby two people can discover what they can do to reduce error in a perception being controlled by another, and in the process reduce error in their own that is influenced by the actions of the other. As an example. Jane disturbs some perception John controls (e.g.his perception of Jane’s contentment) by saying “I’m cold. Could you shut the window”. I call that statement a “display” by Jane. It disturbs John’s perception of her contentment. He acts to bring that perception nearer its reference level by closing the window. The error in Jane’s temperature perception, and thus her comfort level perception, is thus reduced. She feels warmer.
> > 
> > I know that’s much to terse to be much help in describing the usual form of protocols. The main process is the discovery by both parties of what perception(s) each other is acting to control by influencing (the initiator by disturbing) a perception controlled by the other. Frequently the controlled perception is of the value(s) of perception(s) held by the other, in a type of protocol I label “Inform” or “Query”. I am the “continuer” in a “Query” protocol at this moment. I am also an initiator in an “Inform”, because I am doing more than just answering. I am acting to control a perception of your understanding of “protocol” that was disturbed by your query.
> > 
> > I don’t know how many major types of protocol should be distinguished. The process, however, is fairly well defined, at least for dyadic (between two persons) protocols. You can get a lot more detail in the two papers on the topic in the PCT Special Issue of Int ,J, Human-Computer Studies (Farrell, Hollands, Taylor and Gamble, 1998 and Taylor., Farrell, and Hollands, 1998). I thought I had uploaded copies of those papers to Researchgate, but apparently I didn’t.
> > 
> > Hope this helps, and is not too dense.
> > 
> > Martin
> 
> > In the usual definitions, a protocol is (or specifies) constraints to which a communicative interaction must conform in order to be perceived as such. Presupposed metacommunication, if you will. Here are two definitions (for diplomacy and for computer intercommunication, respectively) from the Random House Dictionary:
> 
> > 1. the customs and regulations dealing with diplomatic formality, precedence, and etiquette.
> 
> > […]
> 
> > 1. Computers. a set of rules governing the format of messages that are exchanged between computers.
> 
> > Thus, the Internet Protocol specifies a number of things (such as IP addresses) to which communications from one computer across the Internet to another must comply.
> 
> > You refer to the baby’s initiating protocol and the mother’s continuation protocol. The word here seems to refer to parts of an instance of communication rather than to the whole. Maybe this is analogous to the definitions 2-5 elided above. They are:
> 
> > 1. an original draft, minute, or record from which a document, esp. a treaty, is prepared.
> 
> > 1. a supplementary international agreement.
> 
> > 1. an agreement between states.
> 
> > 1. an annex to a treaty giving data relating to it.
> 
> > In these cases, it’s not synecdoche–the part taken to represent the whole–but rather the word for the whole (‘protocol’) is used to refer to a part. The analogy breaks down because these are parts of a particular instance of protocol-compliant communication (negotiation of a treaty, etc.), and your terms initiating protocol and continuation protocol are at a higher level of generalization. An initiating protocol is any observed behavior by party A that is [intended to be] perceived by party B as initiating an an exchange compliant with a known protocol; and so on.
> 
> > There are two other dictionary definitions which I believe are a bit too divergent to be relevant:
> 
> > 1. Med. the plan for carrying out a scientific study or a patient’s treatment regimen.
> 
> > 1. Also called protocol statement, protocol sentence, protocol proposition. Philos. a statement reporting an observation or experience in the most fundamental terms without interpretation: sometimes taken as the basis of empirical verification, as of scientific laws.
> 
> > > /Bruce
> 
> > On Fri, Dec 15, 2017 at 1:49 PM, Martin Taylor [mmt-csg@mmtaylor.net](mailto:mmt-csg@mmtaylor.net) wrote:
> 
> > [Martin Taylor 2017.12.15.13.14]
> 
> > > [From Rupert Young (2017.12.12 09.30)]
> 
> > > …Do we have anything that puts language (textual) in the PCT framework? …
> 
> > I do, but since it is a substantial part (roughly 100 single-spaced pages out of 600 at the moment) of the book I am writing, I don’t think I can post it, though I might possibly be able to provide drafts to individuals who have the background and who would be prepared to offer helpful comments. All the same, I think it not unreasonable to post a small section that is a sort of introduction to the language part. Here is a PCT “just So Story” about how a baby produces and mother learns a first word.  
> > The working title of the book is “_Powers of Perceptual Control: a PCT enquiry into Language, Culture, Power, and Politics_”, pun intended. My LCS IV chapter is more on the mechanism of protocols, and not really relevant to the question. However, some of my old Layered Protocol papers and chapters might be relevant, since LPT is just a special case of PCT invented before I knew of PCT. A few are on researchgate.
> > 
> > Martin
> 
> –
> 
> Richard S. Marken
> 
> "Perfection is achieved not when you have nothing more to add, but when you  
> have nothing left to take away.â€?  
> –Antoine de Saint-Exupery
