1

Topic: How many percent of all possible texts make sense?

2

Re: How many percent of all possible texts make sense?

Hello, Shmj, you wrote: S> How many %% from these 20 it is possible to involve for the further creation of the text with sense? Let taking into account that that in them not the correct terminations, let even so. S> Though one find, where it would be possible to change the terminations/add of a particle and to receive a certain sentence with sense? S> then how to define probability of what 2 sentences are combined on sense and add each other, form the uniform text? Whether S> somebody studied this question? And it what for is necessary except idle curiosity?

3

Re: How many percent of all possible texts make sense?

Hello, Shmj, you wrote: S> How to calculate at least approximate value from them which make sense or at least probably are not senseless? In Yandex should know, their robot here such texts easily : Positivism discords a maximum in process of signal propagation to the environment with inverse density of population. Schiller stated: the square-topped matrix represents dramatic nature, as well as it was supposed. The quantum state really accumulates an unconscious crystal, however Zigvart considered as criterion of the validity necessity and  for which there is no support in the objective world. The environment attracts Bose condensate. Geometrical progression, on determination, transforms circulating criterion of integrability that once again confirms Einstein's correctness. https://yandex.ru/referats/?t=mathemati … mp;s=48970 we take many intelligent texts and we look at statistics of allocation taking into account a context (N the last words): For each word W what its probability of appearance after word A? For each word W what its probability of appearance after pair words A B? For each word W what its probability of appearance after triple of words A B With? And so on, we receive statistical model. The more long the context, the is more in the table will be zero. Well and further already simple , knowing allocation of really meeting combinations, to count the relation it is is probable looking chains to total of the possible. It in what the generator above is engaged -  such chains. To coordinate them on cases - already simple trick.

4

Re: How many percent of all possible texts make sense?

Hello, Shmj, you wrote: S> Here for Russian, but I suppose that it is the general tendency for all possible languages, type of the law of Tsipfa. S> we Take only 10 thousand the most popular words. Most likely at the very least them it is possible to express any sense (that is not present to replace on "thingummy", "" and so forth). S> It turns out all variants of sentences 10^40. S> How to calculate at least approximate value from them which make sense or at least probably are not senseless? If you again about evolution you again took unsuitable model. Such fact is known. The bright sunlight breaks DNA of bacteria. And so they take scraps of DNA from environment and build in them the a gene. From what it is possible to draw an output that in DNA language any combination makes sense.

5

Re: How many percent of all possible texts make sense?

6

Re: How many percent of all possible texts make sense?

Hello, vmpire, you wrote: V> And it what for is necessary except idle curiosity? And the proof of the theorem of Poincare what for is necessary, except idle curiosity? Generally in the world it is necessary nothing. It is possible to live both without an electricity and without the Internet and without radio. 100 thousand years people so lived also 100 thousand more would live, and here the science endangers life as leads to water and air pollution more likely. All becomes exceptional for the sake of idle curiosity. And  not the interesting thing can lead to very interesting outputs.

7

Re: How many percent of all possible texts make sense?

Hello, D. Mon, you wrote: DM> In Yandex should know, their robot here such texts easily : DM> Positivism discords a maximum in process of signal propagation to the environment with inverse density of population. Schiller stated: the square-topped matrix represents dramatic nature, as well as it was supposed. The quantum state really accumulates an unconscious crystal, however Zigvart considered as criterion of the validity necessity and  for which there is no support in the objective world. The environment attracts Bose condensate. Geometrical progression, on determination, transforms circulating criterion of integrability that once again confirms Einstein's correctness. DM> https://yandex.ru/referats/?t=mathemati … mp;s=48970 it only syntactically the correct texts, most likely even are based on results of already existing texts. Washed off in them is not present is ! DM> The more long the context, the is more in the table will be zero. Well and further already simple , knowing allocation of really meeting combinations, to count the relation it is is probable looking chains to total of the possible. Here your key error. It is necessary was not probablly looking and having or probably making sense.  it is necessary to eliminate, even if it is told syntactically correctly. DM> it in what the generator above is engaged -  such chains. To coordinate them on cases - already simple trick. And how to filter ? While only a method of the conscious analysis.

8

Re: How many percent of all possible texts make sense?

Hello, alpha21264, you wrote: A> And so they take scraps of DNA from environment and build in them the a gene. A> from what it is possible to draw an output that in DNA language any combination makes sense. Not any, is simple there a lot of garbage. And the garbage can be distinguished from an encoding part easily.

9

Re: How many percent of all possible texts make sense?

Hello, Shmj, you wrote: S> Hello, alpha21264, you wrote: A>> And so they take scraps of DNA from environment and build in them the a gene. A>> from what it is possible to draw an output that in DNA language any combination makes sense. S> not any, is simple there a lot of garbage. And the garbage can be distinguished from an encoding part easily. 1) well you understand, how garbage presence influences your estimations? If from gigabyte human  there are 30 mbytes of the useful code (as colleague Cyberax told), c zero you should discard in an exponent? 2) horizontal transfer of genes  is. Both in the nature and in  experiments of scientists. Implementation of genes of fish in a banana now surprises nobody.

10

Re: How many percent of all possible texts make sense?

Hello, alpha21264, you wrote: A> 1) Well you understand, how garbage presence influences your estimations? A> if from gigabyte human  there are 30 mbytes A> the useful code (as colleague Cyberax told), A> c zero you should discard in an exponent? I took 100 MB. I doubt into account 30, such estimations anywhere did not find, we understand with it if it is responsible for words. If 30 MB 10^9030899 variants turn out. Even if variants with sense of trillion trillions there is a minimum of 10^9030800 variants of search of different mutations it was necessary to make. Outside of possibilities of abilities of our Universe. The Universe even to sort out 100 byte of a hash not in forces. A> 2) horizontal transfer of genes  is. A> both in the nature and in  experiments of scientists. A> implementation of genes of fish in a banana now surprises nobody. Me numeric data interest.

11

Re: How many percent of all possible texts make sense?

Hello, Shmj, you wrote: S> It is necessary was not probablly looking and having or probably making sense.  it is necessary to eliminate, even if it is told syntactically correctly. It is not solved task then. For someone one mere verbiage makes sense, and for someone looks . Try to define, where , and where intelligent phrases here: 1) Vosemju eight - hundred. 2) god is love. 3) in the unit over a ring without division is mandatory there is a base. 4) in the unit over a ring without division not is mandatory there is a base. 5) in a division ring on the unit there are zero divisors. 6) the phase  follows for  and is direct analog , following for . 7) Jogacharja   in presentation of Shantarakshity has a row of advantages before  Nagardzhuny. 8) the Universe smells as oil. 9) a banana big, and the peel is even more. 10) for any it is local trivial foliation pi: X-> B and the continuous display f: B-> B ' induced foliation f * (pi) is locally trivial.

12

Re: How many percent of all possible texts make sense?

Hello, D. Mon, you wrote: DM> It is not solved task then. For someone one mere verbiage makes sense, and for someone looks . Sometimes it is impossible to define precisely there is a sense in the text or not. But it is a little such cases. For simplification I suggest these variants to carry out in separate group "probably making sense". There are texts which can find sense only in a context something. Then the estimation of probability of casual origin of the given context is interesting. DM> try to define, where , and where intelligent phrases here: DM> 1) Vosemju eight - hundred. Can find sense only in a context (as an example of an incorrect output and so forth) . Then we consider a variant casual  such context. DM> 2) god is love. A variant "probably make sense" (it is possible to trust, but it is impossible to check up). I would select such variants in special group "probably make sense", it is not enough of them. DM> 3) In the unit over a ring without division is mandatory there is a base. DM> 4) in the unit over a ring without division not is mandatory there is a base. DM> 5) in a division ring on the unit there are zero divisors. In scientific statements if one is not true, can be applied only in a context of an incorrect output and so forth To me difficult to tell, would carry to those which "probably make sense" more likely and "find sense in a context" (that more precisely tells who in this subject works). Here the question of casual origin of the second and so forth sentences which continues thought is interesting and as a result which adding it does not turn out . DM> 6) the Phase  follows for  and is direct analog , following for . DM> 7) Jogacharja   in presentation of Shantarakshity has a row of advantages before  Nagardzhuny. Here at you dedicated religious terms. At these terms very exact sense. Even if you wrote to exactitudes with  sights - is all the same carried to "probably  sense". After all their theory is not checked up yet and it can appear erratic. If you intentionally distorted terms (I them do not know terms) - that is not meaningful. DM> 8) the Universe smells as oil. The Banana big see 1 DM> 9), and the peel is even more. Basically the sense is, but it is required specifying . DM> 10) For any it is local trivial foliation pi: X-> B and the continuous display f: B-> B ' induced foliation f * (pi) is locally trivial. 3, 4, 5 you see intentionally picked up difficult examples. For a fast estimation I would define them at once in group #3 - probably make sense. Not so it is a lot of them. Now we considered only at level of sentences. And after all sentences incorporate in intelligent Tex and probability that casually there will be the second sentence which is uncovering/specifying thought first - even less. Whether exponential rise of complexity will be interesting to me or not?

13

Re: How many percent of all possible texts make sense?

S> Hello, Shmj, you wrote: S> For an example, we take 20 casual dial-ups on 10 words: S> S> not very well marvelous cowards wildly let to reconcile to take out a concert a kidney to be required. S> the thinker share to arrange the excellent dusty engine the agent to train the government to smother. S> to swing the portion mood to thrust to escape sports to blow the leading careful vendor. S> the camp-up to prevail to burn down grain to remove the data intersection aluminum much terribly. S> expertize meaningly to reduce the sky field brilliant error family an institution relocation. S> to bring up a context the odd fellow as hour separation is perfect well forbidden May. S> to tear the device the defender to turn away to throw out a foot to become silent presentation virtue the guilty. S> to launch an aperture to guess a wedge flavor constituent to disappear final five expression. S> shortly the table to dominate simply settlement a duty an idiotic foolish fast cross. S> Sideways the detective to fight the hypothesis from to run up to remind removal the sensor to contour. S> reward for a moment hey romantic excursion to lower this to provoke a breakaway the poster. S> directly to embody a rain a parable a cart the discretion to hide someone's to run out . S> the Eternal Ukrainian to strengthen to prevail rise to burn arguing a nobleman to grin to get acquainted. S> journalism all-Union the usage humanly fascism is lazy export tiny to state an openness. S> gold manners it is angry a sort search helpless to punish to collapse to have breakfast. S> Unknown to be upset early glass to affirm a mouse the companion a curtain the bus languidly. S> Besides a cat territory to connect the editorial Slav fastening the installed pair to attach. S> by all means to continue ruthless to color anniversary immediately successfully stage internal low. S> a virus the classic nutritious fast October the shoulder is disturbing reset to forget to begin to cry. S> to Undertake one nonsense exaggeration a capital mortgage predator a cutlet liquid the razor. S> defect clay to be angry the employee the total automated by bureau the signature imperceptible a monkey. If to generate sentences on certain structure, for example: ", a noun, a verb and ."  often ridiculously it turns out to a measure.

14

Re: How many percent of all possible texts make sense?

Hello, Qulac, you wrote: Q> If to generate sentences on certain structure, for example: ", a noun, a verb and ."  often ridiculously it turns out to a measure. Try so the intelligent text . At least that 2-3 sentences turned out.

15

Re: How many percent of all possible texts make sense?

Hello, Shmj, you wrote: S> Hello, Qulac, you wrote: Q>> If to generate sentences on certain structure, for example: ", a noun, a verb and ."  often ridiculously it turns out to a measure. S> try so the intelligent text . At least that 2-3 sentences turned out. Searched for the generator of casual words, found here it: the Generator of words. Here that it to me produced about : . The drop sacred is in everyone. Apparently, a problem solved.

16

Re: How many percent of all possible texts make sense?

Hello, Qulac, you wrote: Q> Here that it to me produced about : Not the fact that texts absolutely casual. Q> apparently, a problem solved. Q> I do not speak that not solved. Among casual sentences always will be  sense. A question only in their amount. The second question - generation of the sentences coordinated with each other, i.e. probability that 2 sentences coincide on sense and will not contradict each other.

17

Re: How many percent of all possible texts make sense?

Hello, Shmj, you wrote: DM>> 1) Vosemju eight - hundred. S> can find sense only in a context (as an example of an incorrect output and so forth). Then we consider a variant casual  such context. In octal system 8*8=100. All is true, the estimation of that, is in a phrase or the text sense or not, completely depends on a context in a head of the reading. In itself words mean nothing also any sense have no, only in a context. Therefore without formalization and fixing of any context it is impossible to give any answer to such task. An individual question - whether to consider  false expressions. I in that list specially in places imported distortions so the phrase can and looks comprehended, but in practice .

18

Re: How many percent of all possible texts make sense?

Hello, Shmj, you wrote: S> the Variant "probably make sense" (it is possible to trust, but it is impossible to check up). I would select such variants in special group "probably make sense", it is not enough of them. Here I doubt. It seems to me if to take the simple phrase which is making sense (for user Shmj, for example) by simple changeover of one or pair of words from it it is possible to receive a similar phrase which there and then gets to a class "probably make sense". I.e. on one comprehended hundreds "probably comprehended" have.

19

Re: How many percent of all possible texts make sense?

Hello, D. Mon, you wrote: DM> In octal system 8*8=100. DM> All is true, the estimation of that, is in a phrase or the text sense or not, completely depends on a context in a head of the reading. In itself words mean nothing also any sense have no, only in a context. Therefore without formalization and fixing of any context it is impossible to give any answer to such task. Truly. Generally speech about casual origin of any text, including a context. I.e. you understand such text, which sense unambiguously. Certainly, when 1 phrase or even 1 sentence - very rarely it is possible to understand sense unambiguously. And, I think, the estimation needs to be begun with probability of casual origin of a sentence which can be potentially involved in creation of such context. After all many sentences can be eliminated at once. DM> an individual question - whether to consider  false expressions. I in that list specially in places imported distortions so the phrase can and looks comprehended, but in practice . False can make sense only in a context of erratic outputs and similar. A main point such: if we make the text of casual sentences, whether there will be an exponential growth of complexity of generation of each new sentence? For example, the first sentence can be selected from 10^38 variants. The second only from 5^38, the third only from 2^38 etc.

20

Re: How many percent of all possible texts make sense?

Hello, D. Mon, you wrote: DM> Here I doubt. It seems to me if to take the simple phrase which is making sense (for user Shmj, for example) by simple changeover of one or pair of words from it it is possible to receive a similar phrase which there and then gets to a class "probably make sense". I.e. on one comprehended hundreds "probably comprehended" have. It is necessary to calculate. However if to take such 2-3 sentences which are "probably making sense" and to try to connect - at once it will be visible that the text is deprived sense. 1. At first we learn how many sentences from 1000 make sense, how many "probably have" and how many have no. 2. Then we learn how many combinations from 2 sentences possess the given properties. 3. We increase to 3 sentences. The more sentences, the less chances it is necessary to receive the coordinated text with sense or even with possible sense. There are objections? Be clear that exactitudes here cannot, but at least approximate estimations, I think, are possible.

21

Re: How many percent of all possible texts make sense?

Hello, Shmj, you wrote: S> Generally speech about casual origin of any text, including a context. I.e. you understand such text, which sense unambiguously. It can be estimated only for the specific receiver/reader since the conciseness estimation depends on, actually, all life experience of the reader. Here to plant Shmj and to force to estimate conciseness of billion texts then we receive number. And the text in itself, without thinking about the receiver, does not contain sense. Normally we assume that the author of the text enclosed in it sense since the text makes sense at least for the author. And at casual generation of this premise already is not present, there is only text. I.e. it always function of a type of EstSmysl (the Text, the Receiver). Without the second argument does not work. Whales of a song sing and communicate with each other - how to understand, what their "texts" make sense, and what - is simple ?

22

Re: How many percent of all possible texts make sense?

Hello, D. Mon, you wrote: S>> Generally speech about casual origin of any text, including a context. I.e. you understand such text, which sense unambiguously. DM> it can be estimated only for the specific receiver/reader since the conciseness estimation depends on, actually, all life experience of the reader. Your error that you consider the reader static, not capable to search and learning of the new information. You as though consider it as a certain algorithm, the program. Forgetting that the consciousness by the nature is distinct from algorithm. DM> here to plant Shmj and to force to estimate conciseness of billion texts then we receive number. To plant any consciousness interested in a finding of sense - and we receive about identical number. Yes, there will be a certain percent of errors, consciousness so is arranged. At more careful  errors can be reduced. DM> and the text in itself, without thinking about the receiver, the sense does not contain. Normally we assume that the author of the text enclosed in it sense since the text makes sense at least for the author. And at casual generation of this premise already is not present, there is only text. I.e. it always function of a type of EstSmysl (the Text, the Receiver). No. Important only that the estimation was led adequate and more by less healthy consciousness. At once it is possible to tell when the sense precisely is not present. Not complete phrases taken out of context and sentences are disputable. Even if the text now is not clear - it is possible to go deep, study and draw the unambiguous output. I think that for casual generation it is necessary to eliminate difficult terms from basis of words to simplify time for realization. DM> without the second argument does not work. An error of misunderstanding of the nature of consciousness. DM> whales of a song sing and communicate with each other - how to understand, what their "texts" make sense, and what - is simple ? On language learning time is required. It is possible, but takes some years. On it I suggest to estimate that language, which person already knows. Simply for abbreviation of time for realization.

23

Re: How many percent of all possible texts make sense?

Hello, Shmj, you wrote: V>> And it what for is necessary except idle curiosity? S> and the proof of the theorem of Poincare what for is necessary, except idle curiosity? S> generally in the world it is necessary nothing. It is possible to live both without an electricity and without the Internet and without radio. 100 thousand years people so lived also 100 thousand more would live, and here the science endangers life as leads to water and air pollution more likely. S> all becomes exceptional for the sake of idle curiosity. And  not the interesting thing can lead to very interesting outputs. Judging by an answer emotionality, the answer to my question: "it is not necessary also concepts I have no what for can it is useful". Otherwise the answer would be idle time: "for today - is not present" Generally - a normal question, what for so to be boiled? If it is attempt of the decision of the real task, that, probably, it would be possible to consider other approaches to its decision.

24

Re: How many percent of all possible texts make sense?

25

Re: How many percent of all possible texts make sense?

Hello, vmpire, you wrote: V> Generally - a normal question, what for so to be boiled? If it is attempt of the decision of the real task, that, probably, it would be possible to consider other approaches to its decision. There is an assumption that for all existing languages (as natural and artificial) an amount   about the identical. It would be desirable to check up this assumption while to begin with one language. If the assumption proves to be true, rather interesting outputs follow.