I also wonder how somebody would watch his cows, while not constantly remembering them. In daydreaming the cows would get simply lost. The cowherd is simply more relexed as the cows would no more damage the fruits. Still he is watching them to get not lost.
That seems to be a possibility, but I don’t read MN 19 as suggesting that the task goes beyond preventing the cows from nibbling the harvest. The imagery in MN 19 looks to me to be pointing to just one task. That task having been completed, I don’t think the sutta suggests that there is a further task to do in preventing the loss of the cows.
An extended description of fourth right effort in Samvarappadhana sutta involves keeping the attention on certain perceptual image (nimitta) or selective recognition (sanna):
katamañca bhikkhave anurakkhaṇappadhānaṃ? Idha bhikkhave bhikkhu uppannaṃ bhaddakaṃ samādhinimittaṃ anurakkhati aṭṭhikasaññaṃ pulavakasaññaṃ vinīlakasaññaṃ vipubbakasaññaṃ vicchiddakasaññaṃ uddhumātakasaññaṃ. Idaṃ vuccati bhikkhave anurakkhaṇappadhānaṃ.
http://awake.kiev.ua/dhamma/tipitaka/2S ... ggo-p.html
Pls forgive my clumsy and hyper-literal translation of AN 4.14 above –
And what, O monks, is the effort to guard/protect? Here, O monks, a monk guards the arisen good “sign of concentration” – the perception of bones, the perception of the worm-infested (corpse), the perception of the blue-black (corpse), the perception of the festering (corpse), the perception of the punctured (corpse), the perception of the swollen (corpse). This, O monks, is called the effort to guard/protect.
It might not be the “recognition” or “apperception” function of sañña at work here, but rather, the imagination/conceptualisation at work. Of the 6 adjectives applied to a corpse in this sutta, 3 (uddhumātakaṃ vinīlakaṃ vipubbakajātaṃ) are shared with MN 10’s discussion on cemetery contemplations. In MN 10, the optative passeyya is used to indicate a hypothetical sort of “sight”, which the meditator then applies/upasaṃharati to his own body; this suggests adhivacanasamphassa, rather than the more direct paṭighasamphassa at work.
The other connection to satipaṭṭhāna in AN 4.14, lies in its reference to samādhinimitta, which MN 44 would define as the 4 satipaṭṭhānas. As I mentioned in my earlier post, the verb upasaṃharati certainly allows for a more discursive and recollective stance, but I do not know how the more subtle pajānāti (discerns) or sampajānakārī (applies awareness) requires recollection.
That being said, my own interpretation would require that even the effort to protect in terms of “just knowing” is mediated by a sankhāra. What does not seem apparent is whether or not one needs to deliberately recall such an instruction, or whether the sankhāra works automatically. After all, SN 12.25 makes clear that not all sankhāras arise sampajana. Kamma can also be done asampajana (unconsciously). No distinction is drawn between good kamma or bad kamma in terms of their potential to be asampajana.
Which brings me to your query -
How could the Bodhisatta steadily see and discern two kind of thoughts, and apply proper efforts, if he didn't maintain the remembrance of what he was doing?
This is an absolutely legitimate point, and it appears from the suttas that the application of the 4 Right Efforts will certainly require one to constantly recall and remember to "do" something at a fairly "coarse" level. But, this still cannot explain whether recollection and "doing" still has much of a role, when the defilements have reached a fairly untroublesome level, such that one simply pajānāti/knows that they are present or absent. This is in fact brought to the fore in the Satipaṭṭhāna suttas' instructions to dhammesu dhammānupassī pañcasu nīvaraṇesu (practise dhammānupassanā with reference to the 5 Hindrances). The "bare awareness" refrain of ñāṇamattā patissatimattā is also applied to the 5 Hindrances, which suggests that one could take a very stable and neutral stance when facing the Hindrances.