Respected Mr. Malfet
Phonosemantics says that every sound has specific meaning assigned by nature. But it never says that every word has specific meaning. It is to be understood that the words are not the true representation of any object. Out of infinite identifications of the object, we select some of them, convert them into vocal gestures and speak out with air. The flow of multiple vocal gestures forms a word. In this way the word has incomplete identification of the object. The object having infinite number of identifications cannot be described within incomplete identifications. For example take a word ‘father’. Father is person; male; can be animal or human; has a biological, psychological, and intellectual entity; has more power relative to female; approves, protects and restricts the children; provides liveliness to the children; provides half of his DNA to the children. We can have a long list expressing properties of father. In Chinese we call /pa/ for father. The semantic value of this /pa/ is “the entity providing approval”. They have selected only one property out of infinite properties. You can explain as why in China father is called /pa/. But you cannot say that the meaning /pa/ is always ‘father’. Any entity which provides approval to anything will be called /pa/. We have long list of entities that provide approval, protection, and bond to any other entity (living and non-living). This list can have many names including elder brother, mother, police, door, arms, cloths, doctor, advocate, night, 100, watchman, medicine, warheads,  ....etc
In this way we cannot make out the entity by its name (composition of phonemes). We can only explain as why the Chinese says /pa/ to father. We cannot say that translation of /pa/ is father.
We have infinite number of images in universe. Nature devises these infinite numbers into finite number of groups. Nature assigns specific phonemes to each group. Once the image is converted into phoneme; it loses its original existence. While reconverting the phoneme into the image, it can represent the properties of its group only, not its original property. Hence the word itself cannot express the originating image of the object.
If you ask me to define any name. It seems still difficult. For example a word /komba/ can be defined like this- Entity of + adjective of /m/ + noun of /b/ + adjective of /o/ + noun of /k/. The semantic values of /m/, /b/, /o/, and /k/ are to be taken from chapter no 5 of the paper. I think there will be no fruitful result will come out. All /m/, /b/, /o/, and /k/ have large number of images. The selection of the image can only be done if you already know the meaning of the word.

The selection of the image can only be done if you already know the meaning of the word.

And so the final death blow has been dealt. There's no coming back from this, unfortunately. Never possible to test, always making the hypothesis fit the evidence. Never a chance to be wrong. I had my suspicions anyway, but I never thought the OP would be the one to completely close the deal on this one.

I have some questions.

(1) Do humans possess this information on phonosemantics somewhere in their minds? If a person's semantic modules (providing we have them, for this example) were switched off, and he was listening to words, would this person, according to your idea, have a subconscious representation of the phonosemantic nature of meaning of the word being said, without knowing what the word means?

(2) What about homophonous words in different languages? If I have my representations for a word in English, and I learn a second language that has a totally different word but happens to sound exactly the same, how can both these words have exactly the same phonosemantic reading? Does anything change when someone learns a word already in their lexicon? Does it unite with the other one? Is a separate yet related mental storage place set up for it?

(3) If these things are something that don't really clearly exist in the human mind (rendering the previous two questions inapplicable) then how is it you have managed to decipher the system?

(1) Do humans possess this information on phonosemantics somewhere in their minds? If a person's semantic modules (providing we have them, for this example) were switched off, and he was listening to words, would this person, according to your idea, have a subconscious representation of the phonosemantic nature of meaning of the word being said, without knowing what the word means?

It is a known fact that “the overwhelming majority of animals are born knowing how to speak their species’s language. It doesn’t really matter where those animals are born or raised, because their speech seems to be mostly imprinted in their genetic code.” [Palmer, Brain. June 26 2012. Health Science & Environment. Do animals from geographically distant areas speak the same language ?] It proves that there must be some universal relationship between sounds and their meanings. The humans are also animals and hence they also possess the phonosemantics somewhere in their existence.
There is no basic difference between animals and humans. The animal interaction is a pure psychological interaction. And in case of humans, they frame up their interaction in such a way so that an intellectual listener can understand the intellectual message.
Phonosemantic is a mechanism by which we convert physical sounds into psychological meanings. In case we switch off the phonosemantics, our hearing and speech both will shut down. Hence there will be no question about the “meaning of word without knowing that”. Here the word ‘knowing’ is not limited to intellectual knowing. It includes ‘psychological knowing’ too. If you hear a sound, it will give you psychological impact. If you say that you are feeling a sound, it means that you are psychologically hearing the sound. 

(2) What about homophonous words in different languages? If I have my representations for a word in English, and I learn a second language that has a totally different word but happens to sound exactly the same, how can both these words have exactly the same phonosemantic reading? Does anything change when someone learns a word already in their lexicon? Does it unite with the other one? Is a separate yet related mental storage place set up for it?

I have number of words in my list, but all in English. Here I am taking a word /bæŋk/ . The phonosemantic meaning of this word is “lively consciousness /ŋk/ of visibility of bond/protection in existence /bæ/”. Now if the existence is ‘water’, the meaning will be ‘river bank’, and if existence is ‘money’, it will be ‘money bank’.  In fact the word ‘bank’ does not denote the ‘money bank’ or ‘river bank’, it denotes “an existence which protects something with alertness”. In this way you can have two entirely different meaning of same word. It is to be understood that while naming any object, we indicate it with shortest possible description or utility.
Please forgive me for my poor English.

It proves that there must be some universal relationship between sounds and their meanings. The human is also an animal and hence they also possess the phonosemantics somewhere in their existence.

You're concluding in a way that no one with a background in scientific methodology would ever conclude. This is speculative conjecturing at best.

Phonosemantic is a mechanism by which we convert physical sounds into psychological meanings.

Well, no. Not really. The mechanism by which we convert sound into meaning can be - and has been - explained by processes that have no recourse to any phonosemanic account. This is a statement that would need to be backed up somehow, not used as evidence to back up another point.

The phonosemantic meaning of this word is “lively consciousness /ŋk/ of visibility of bond/protection in existence /bæ/”.

Lively consciousness? That doesn't seem like a reasonable/likely category. What rationale do you have for this? Never in a million years would I see that word and come to that conclusion. Is my phonosemantic machine turned off?

In fact the word ‘bank’ does not denote the ‘money bank’ or ‘river bank’, it denotes “an existence which protects something with alertness”.

Does it? I don't think it does. You'd need to convince me of this.

You might want to check out Hockett's Design Features of Language (very famous piece of work in Linguistics) and his categorisation of the different systems of communication between animals and humans.

In this way we cannot make out the entity by its name (composition of phonemes). We can only explain as why the Chinese says /pa/ to father. We cannot say that translation of /pa/ is father.

What you are describing is a word game. If it brings you pleasure, I hope you will continue to play it. However, if you wish to be understood as participating in science (and, specifically, if you wish to be taken seriously by Euro-American linguists), you will have to reformulate things in the manner I described.

Whichever you decide, good luck!

Isn't an army or soldier or bomb a better example of something that protects something with alertness? Why are these not pronounced as "bank"?

You are very clearly telling stories rather than doing science. Stories are great. But they don't explain how languages work.