[Physics] Physics list vs physics-announce

Tue Oct 18 23:19:34 CEST 2016

Arend,

I am a physicist and inventor.  I spent half my working career programming computers.  I also took a course in logic at Western Reserve University as an undergraduate.  I wrote computer code for the first computer to analyze human speech in the mid-70's.  So I think I can speak with reasonable authority on this subject.  

There are AI programs out there that can understand human speech and "talk" to a human in our language.  These very large computer AI programs are currently installed on very large computers, but in time they will appear in household products.

The logic program that I proposed would be considerably less sophisticated than a full AI program that understands human speech.  The logic program only needs to divide each sentence into parts of speech, similar to what students do in 8th grade English class.  The program would then determine if these parts of speech work together logically. . . a relatively straight forward procedure.  The program would have some difficulty with slang expressions; but dictionaries of slang expressions could be installed in the program.  Also, normally in printed text, there would be less slang used.  

A logic program would be somewhat less difficult to write then a speech translation program.

Sincerely,

Maurice Daniel, Master Inventor
mdaniel at masterinventor.com

On Oct 18, 2016, at 2:59 PM, Arend Lammertink wrote:

> Hello Maurice,
> 
> It would definitely be nice if computers could be able to assist in the kind of tasks you envision.
> 
> However, it is incredibly difficult to analyze natural languages with a computer. Computers are 100% logic. It's literally built up completely of binary logic. From that base, it is possible to implement arithmetic. This makes it a very useful machine in the field of science and engineering and it has been for decades already. 
> 
> When computers became more powerful in the 90's, it became possible to use it for music and even video. Other tools include word processing, e-mail. etc.
> 
> The fundamental problem with computers is that, as far as I know, we still have incredible problems with expressing "meaning", for instance of a word, into a form which can be handled by a computer.
> 
> If you translate a text with Google translate, for example, you will find that these are not very good. They can give you a pretty reasonable idea about what is being said in another language, but you miss a lot of detail. And that's only the translation from one language to another. 
> 
> For translating one language to another, the computer does not have to "understand" the word it's translating. Sure, one can argue that associating, for instance the word "hello" in English with "hallo" in Dutch, but that's not "really" understanding. 
> 
> So, what I mean to say is that "machine translating" is a lot easier to do, because the computer only needs relatively little "understanding" of what is being meant by the words being translated. 
> 
> What this means is that computer science, as far as I know, is still years away from being able to analyze  the "meaning" of natural languages. So, when the time has come that machine translators can accomplish a quality comparable to a human translator, computer science will probably also begin to be able to do the kind of text analysis I think will be needed for the type of proposal you envision. 
> 
> To cut a long story short: It would be very nice if computer science were able to analyze human languages, which would be needed to do this. Progress is definitely being made on this subject but, as far as I can tell, we're still a few years away before ideas like yours can be accomplished. It seems very simple at first glance (I've thought about building a translator a long time ago, too :) ), but it's very challenging to cover "meaning" with machines that can actually do little more than binary logic. 
> 
> Best regards,
> 
> Arend.   
> 
> 
> 
> 
> 
> On Tue, Oct 18, 2016 at 6:25 AM, Master Inventor <mdaniel at masterinventor.com> wrote:
> Arend,
> 
> I am sending you an essay I wrote a short time ago that may be of great interest to you.  I do not know if you have the resources to write the computer program I describe in the essay, but perhaps you know a university that would be willing to take on the task.  It would certainly make it easy to sort out what papers to publish without having to pass judgment on them. 
> 
> ---Maurice---
> 
> <Maurice Photo, Small Size.jpg>
> 
> 
> 
> Maurice Daniel, Master Inventor
> mdaniel at masterinventor.com
> 
> 
> Logic Analysis Application Software
> Maurice Daniel
> August 4, 2016
> September 10, 2016
>  
>  
>             It is said that all things are possible; but when it comes to physics this statement is simply not true.  But it is difficult to set limits on the unknown.  As scientists we must always have an open mind, but we cannot believe everything, we must separate the possible from the improbable and the nonsensical.  The peer review process attempts to do this, but instead it completely stifles creativity; it accepts new knowledge in very small increments and only after an excruciating review process lasting inordinate amounts of time.  The internet accepts nonsense with open arms and uses it to berate scientists for their inability to see the obvious.  There seems to be no middle ground where new ideas in physics can be discusses without prejudice to see where they lead.  Initial ideas are often far from the mark, but if they have perhaps a kernel of truth they can be followed into productive areas never before explored with new eyes. 
> 
>             The body of scientific thinking certainly contains many errors, but it can not be ignored.  It also contains countless observations and measurements that must be accounted for in any new theories.  So there are two rigorous tests that new theories must pass.  They must explain the class of observations pertaining to their subject area and they must pass the test of logic.  New theories may supercede old theories but they must have predictive power that at least equals the old theories. 
> 
>             There are tools that can be used to better evaluate new speculations.  One method is to prepare an exhaustive list of all possible explanations and then to evaluate each of them to arrive at the most likely explanation.  Always the last explanation on the list must be “none of the above”.  Unfortunately, in the rush to fame and glory the systematic procedure of evaluating alternatives is often simplified to lip-service resulting in explanations that fail the most basic test of logic.  This seems to be particularly true in the field of astrophysics where core observations are explained by layers upon layers of pure speculation.  As a result every major theory of astrophysics is contradicted by observation.  (These contradictory observations are brushed aside so as to not confuse the public and other astrophysicists as well.) 
> 
>             However, there is one tool available that can quickly weed out nonsense from legitimate speculation; this is the tool of logic.  Logic is a minor branch of mathematics and it is also one of the essential classes taught in the philosophy curriculum.  The tools of logic are quite sophisticated.  A paragraph of text can be broken down into a set of logical relationships without knowing the subject matter of the paragraph.  Narrative text can be broken down into a set of assumptions (or facts) and conclusions connected by assertions (logical operators).  No matter how complex the text the rules of logic can be applied to untangle the statements and determine if any errors were made in reaching the conclusions.  This test of logic could easily be applied to the text by a computer.  (The internal hardware and software of a computer have logic circuits built in as a basic part of its operation.)  It would not be very difficult for a computer software program to be devised that could analyze paragraphs or pages of text, identify the assumptions and conclusions, and then identify if there are any errors in logic.  The program could also print out a list of the assumptions and a list of the conclusions. 
> 
>             It would be a great service to the scientific community if such a logic analysis program were widely available to the editors of scientific publications.  For one thing this could reduce or eliminate the need for peer review of scientific articles. The list of assertions and “facts” identified by the program could quickly be given a reality check by editors, readers, and authors.  Articles that failed the test of logic would be sent back to their authors for corrections.  Publications of pure scientific speculation could then be published without the fear of publishing nonsense. 
> 
>             A logic software application of this type could be added to all word processors, similar to the spell-check module found in most word processors.  This would allow authors to quickly check their own work for logic errors before they publish it on-line and make a fool of themselves.  As a stand-alone software application it could be used to scan one-line information, news, political statements, advertisements, and other articles for errors in logic (which often occur).  The wide availability of this type of software application would greatly reduce conflict and miscommunication among all peoples. 
> 
> ---Maurice---
> 
> 
> 
> _______________________________________________
> Physics mailing list
> Physics at tuks.nl
> http://mail.tuks.nl/cgi-bin/mailman/listinfo/physics
> 
> 
> _______________________________________________
> Physics mailing list
> Physics at tuks.nl
> http://mail.tuks.nl/cgi-bin/mailman/listinfo/physics

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.tuks.nl/pipermail/physics/attachments/20161018/999d3e6b/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Maurice Photo, Small Size.jpg
Type: image/jpeg
Size: 17131 bytes
Desc: not available
URL: <http://mail.tuks.nl/pipermail/physics/attachments/20161018/999d3e6b/attachment-0001.jpg>